CAMBRIDGE CLASSICAL STUDIES

General editors THE ROMAN BAZAAR

R. L. HUNTER, R. G. OSBORNE, M. MILLETT,
A Comparative Study of Trade and Markets in a Tributary Empire
D. N. SEDLEY, G. C. HORROCKS, S. P. OAKLEY,
W. M. BEARD

PETER FIBIGER BANG

CAMBRIDGE
UNIVERSITY PRESS
 CAMBRIDGE UNIVERSITY PRESS
Cambridge, NewYork, Melbourne, Madrid,CapeTown.Singapore, SaoPaulo, Delhi
Cambridge University Press CONIUG1 OPTIMAE AC FII. IIS CAR1SSIM1S
The Edinburgh Building, Cambridge di2 SRV. UK
PublishedintheUnited StatesofAmericabyCambridge University Press,NewYork
www.cainbriclge.org
Informationonthistitle:www.cambridge.org/978o52l855327
e Faculty ofClassics, University ofCambridge 2008 Some valuing those of their own side or mind,
Still make themselves the measure of mankind:
Thispublicationisincopyright.Subjecttostatutoryexception Fondly we think we honour merit then,
andto theprovisions ofrelevant collective licensing'agreements,
no reproduction of any part may take place without When we but praise ourselves in other men.
the written permission of Cambridge Universitv Press. Alexander Pope, An Essay on Criticism
First published 2008

Printed inthe United Kingdom attheUniversity Press,Cambridge

A cataloguerecordforthispublicationisavailablefromtheBritishLibrary
Library of Congress Cataloguing in Publication data
Bang, PeterF. (Peter Fibiger)
TheRomanbazaar/ PeterPibigerBang.
p. cm. - (Cambridge classical studies)
Includes bibliographical references.
ISBN 978-0-521-85532-7 (hardback)
I. Bazaars(Mmtets) - Italy- Rome-'History: 2.Rome(Italy) - Commerce - History.
3. Rome (Italy) - Economic conditions. I. Title. II. Series.
HI.-5474. I82R663 2008
381 . I -clc22 2008031924
ISBN978-0-521-85532-7hardback

EX L16RIS
UNIV6RSITAT18
WVIOMAQENSIS I

Cambridge University Presshasnoresponsibility forthepersistence 01-

accuracy of URL.s for external or third-party internet websites refeiTed to
in this publication anddoes not guarantee thatanycontent on such
websites is, or will remain, accurate orappropriate.
 CONTENTS

List of figures and tables page viii

A note about abbreviations x

Preface Xl
Key dates and terms in Mughal history Xlll
Roman weights, measures and coins XV

Prolegomena

PART I THE ROMAN EMPIRE AND THE COMPARATIVE

STUDY OF PRE-INDUSTRIAL SOCIETY 15

i Beyond The Ancient Economy"! Trade in the Roman

Empire and the problem of comparative history 17
2 An agrarian empire between market and tribute:
situating interregional trade in the Roman world 61

PART II IMPERIAL BAZAAR 129

3 A rough trading world: opaque, volatile and

discontinuously connected markets 131
4 A thin line: portorium, protection and predation 202
5 Community: cult, courts, credit and collaboration in
the bazaar 239
Epilegomena: taking stock - the world of goods 290

Bibliography 307
Index 351

Vll
 LIST OF FIGURES AND TABLES

S.I d, bottom, floor mosaic from the section of the

FIGURES AND TABLES
Frontispiece ThebazaarintheoldtownofDamascus onedayin
the spring 2001.
Figures
i. t Olusegun Obasanjo is sworn in as president of
Nigeria (PolFot, Denmark) page 39
2. 1 The Mughal Empire (J. F. Richards, The Mughal
Empire, The New Cambridge History of India I, 5,
Cambridge, 1993, p. iv)
2. 2 Rough model ofagrarian economy
2. 3 Roman GDP, surplus and tributary empire
3. 1 Grain prices in Amsterdam and Danzig in the
eighteenth century (J. de Vries andA. Woude,The
First Modern Economy. Success, Failure and
Perseverance ofthe Dutch Economy, 1500-1800,
Cambridge, 1997, p. 418) 148
3. 2 Roman Egypt and the province ofArcadia 159
3. 3 Price ofwheatin fifth-century province ofArcadia 163
3.4 Price ofbarley in fifth-century province ofArcadia 164
3. 5 Price of wine in fifth-century province of Arcadia 167
5. 1 Piazzale delle Corporazioni (plan from CILXIV, 251-252
suppl. Ostiense, p. 662)
5. ia, top, view ofPiazzale delle Corporazioni from
the theatre in Ostia
5.i b, bottom, plan ofthe complex
5.ic, top, close-upof the small temple in the centre
of the complex
portico belonging to the navicul(arii)
et negotiantes Karalitani (C1L XIV, 4549, 21 =
no. 21 on the plan)
5. 2 Palmyra and its trading world 255-257
5. 2a, Map of the central part ot'Palmyra's trading
network (from G. Young, Rome 's Eastern
Trade. International Commerce and Imperial
Policy, 31 BC-AD^o^, London, 2001, p. 140)
5. 2b, top, view of the central cella (to the right) and
the courtyard of the enormous temple of Bel in
Palmyra
5. 2C, bottom left, section of the Bel temple's portico
5. 2d, bottom right, close-up of one of the statue
bases with bilingual inscription
81 Where no source is indicated, the graphics and photographs are
the author's own.
85
114
Tables
2. 1 Estimate of disposable surplus, state impact and
interregional trade in the second century AD 87
2. 2 Living standards for different groups of population
under different productivity regimes I 18
3. 1 Variation in price levels in Arcadia 161
3. 2 Commodity prices, probably in Oxyrhynchus, during
fifteen months of the fourth century AD 168

vni
IX
 A NOTE ABOUT ABBREVIATIONS
A few minor and unproblematic deviations apart, in referring to
ancientsourcesI usethestandardabbreviations oftheOxfordLatin
Dictionary, Liddell and Scott's Greek-English Lexicon, 9th edn.
andJ.Gates etal., Checklist ofGreek, Latin, Demotic andCoptic
Papyri, Ostraca andTablets.
PREFACE
"Habent sua fata libelli" the Roman writer remarked in days of
old. This is just as true of scholarship as of literature. Over the
years I have beenmet with extensivegenerosity, support, patience
and liberal-minded curiosity from many people and institutions.
I am delighted to offer my heartfelt thanks to all of these at the
completion of this study. It originated as a PhD submitted at the
Faculty of Classics in Cambridge in 2003. The Danish Research
Agency financed the three preceding years of work in Cambridge,
for which Corpus Christi College provided a wonderful setting,
both for me and my family. Fondation Idella, Knud H0jgaards
Fond, Corpus Christi College and the Faculty of Classics gener-
ously gave supplementary funding. In Denmark, Erik Christiansen
and Uffe 0stergard helped me lay the foundationsto the research
I conducted in Cambridge and have been unwavering in their sup-
port. In Leicester, David Mattingly tolerated a polemical visitor
and provoked me to conceive the first ideas for a comparative
study of the Roman Empire. Sarah Scott encouraged me to think
of Cambridge. In 2001, The Danish Institute in Damascus and its
then director, Peder Mortensen, made a research trip to Syria and
the ruins of Palmyra an unforgettable experience. I owe particular
thanks to Joyce Reynolds, who has been such a kind friend, as
well as Dorothy Thompson and Michael Sharp for discussing mat-
ters ofepigraphyandpapyrology with me. RichardDuncan-Jones
and Muzaffar Alam gave me kind advice on different occasions.
Michael Crawford kindly let me have a preview of the new edition
of the text for the Asian Customs Law, still in preparation.
Richard Saller generously invited me to Chicago, supervised
my research for a term, did much more than one could reasonably
expect and introduced me to Walter Scheidel, who has continued to
offer friendship and inspirational discussions over the years. Keith
Hopkins volunteered to supervise my second year and made it one
Xl
 PREFACE

of intellectual play and expanding horizons. Chris Bayly took a

kind interest in the project from very early on, gave me invaluable
guidance on Indian history and has grown into a trusted colleague KEY DATES AND TERMS IN MUGHAL HISTORY
and collaborator. Vincent Gabrielsen and my other colleagues in
the Saxo Institute at the University of Copenhagen had faith in
me, took me on as lecturer before the completion ofmy research
and encouraged my work in every aspect. However, with that
also came new obligations and new projects. Kari Erik Frandsen. 1526: Battle of Panipat, foundation of Mughal
the former director of the Institute, gave me the period of peace rule in northern India.
and quiet which enabled me to return to Cambridge for a second 1526-30: Babur, first Mughal emperor.
round in the first halfof 2007 andto start turning the thesis into a 1530-56: Humayum, reign interrupted by exile
book ThistimeRobinOsbornekindlyextendedhospitality aswell from 1540 until victorious return shortly
as stimulating discussions at King's, while the Danish Research before his death.
Council gave financial support. Dick Whittaker and Greg Woolf 1556-1605: Akbar.
originally examined the thesis, offered penetrating observations
andmadethedefence anoccasion whichI remember withjoy. The Conquest and cooptation of the Rajputs, conquest of Gujarat and
Bengal.
comments oftwo anonymous readers, aswell asmy lecturership
committee, further helped me to improve the manuscript and the 157^-85: seat of government transferred from
argument. Agra to FatehpurSikri, a newly erected
Above all, however, my thanks go to Peter Gamsey who first capital and palace complex, which was
took me on at Cambridge. I can hardly believe almost ten years again abandonedafter only a few years
have passed. Enjoying the privilege ofhis inspirational guidance, of use.
firm support, loyal friendship, tolerant patience and incisive com- 1605-27: Jahangir.
ments has not only benefited my work immensely, it has also set 1627-57: Shah Jahan.
anexample to whichI canonly everaspireinmy ownteaching.
On the domestic front, I owe an enormous debt of gratitude Expansion into the Deccan.
to my parents, who taught me to go my own ways, and to my in- c. 1632-43 Taj Mahal: construction of the mausoleum to the
laws;theyhavesupportedmeineverywaypossibleovertheyears. favourite wife of Shah Jahan, Mumtaz
Ultimately, however, it was my family, Helena, Julian and Mette. Mahal.
whomadeit allpossiblebyagreeingtofollowmeouton"foreign 1639-48: construction of the Red Fort of Delhi
adventures", bytoleratingmywhimsandbygivinglovingsupport which was henceforth elevated to serve
in times ofagonising doubts. It is to them that I dedicate this work. as the imperial capital under the name:
I wish it were perfect, they would deserve nothing less, but this is Shajahanabad.
the best I could do.
1650-56: construction of the Jami' Masjid
mosque in Shajahanabad/Delhi.
Copenhagen,autumn 2007 1658-1707: Aurangzeb (Alamgir).
1685-87: Annexation of the tributary sultanates of
Bijapur and Golconda in the Deccan.
Xll
Xlll
 KEY DATES AND TERMS IN MUGHAL HISTORY

Lastthirdofthecentury:insurgencyoftheHinduMarathasdrags
downthe Mughals in a fruitless warof attrition.
ROMAN WEIGHTS, MEASURES AND COINS
Gradual weakening anddisintegration of Mughal rule follows in
the decadesafterthe deathofAurangzeb.
1707-12: Bahadur Shah.
The principal grain measures
I7I2-I3: Jahandar Shah.
I7I3-I9-- Farrukhsiyar. I modius = c. 6. 55 kg or c. 8. 62 1
17^9-48: Muhammed Shah.
[ medimnos = c. 40 kg or 52. 53 1
1739: occupation of Shajahanabad (Delhi) by I artaba = c. 30. 2 kg or c. 391; in late antiquity 21. 5 kg or 29 1
the Persian ruler Nadir Shah. The
Mughal capital is plundered and the Liquid measures
famous Peacock Throne carried away to
Persia. End of effective Mughal power. i metrete 39. 39 1
1837-57: BahadurShah2 (deposedafterthe i kotyle c. 0. 25 1
Great Indian Mutiny). I monochoron c. 7 1
i sextarius = c. \ pint (16 sextarii to a modius)
Mansabdar is the Mughal term for the individual members of
the imperial aristocracy. The mansab refer to the ranks which the Square measures
nobles were awardedby the emperor.
Jagirdar is occasionally used instead oiMansabdar. However. I mgerum = c. 2, 500 m square
heretheemphasisisontherevenueSissigments, jagirs, whichthe i aroura = c. 2, 760 m square
nobles received inreturn forservingthe'emperor. Coins
Zamindaris usedto denote local landowningelites acrossthe
Mughal Empire andassuch covers a broad spectrum ofwealth, i denarius (silver) = 4 sesterces (HS, bronze) = 16 asses
statuses and influence.
i aureus (gold coin) = 25 denarii = 100 sesterces
myriad denarii: in late antiquity, the silver contents of the denarii
dwindled to almost nothing and its value correspondingly col-
lapsed. Hence we now find the notion of myriad denarii (10, 000
denarii)

solidus: in the same period a new gold coin, the solidus, from the
time of Constantine struck at 72 to the pound, became the new
backbone of the Roman imperial coinage.
i drachma = 6 obols (in Egypt 7 obols)
i mina = 100 drachmai
I talent == 60 minai = 6, 000 drachmai
XIV
XV
 PROLEGOMENA
an Oriental "artefact", as the form of trade characteristic of the
pre-modem high cultures of the agrarian world. The bazaaris not
incompatible with a Roman setting, far from it; in Damascus it is
still possibletoobservehowthetwoco-exist,theoneorganically
built into the other, in the narrow alleys leading from the old Via
Recta to the Umayyad Mosque, the former temple of Jove (see
Frontispiece).
Evidently, part of the problem of ancient economic history
is the poor and fragmented state of the evidence. Compared to
historians ofEuropean tradeandcapitalism from theMiddle Ages
through to the industrial revolution, the situation ofthe ancient his-
tonan is deplorable. Nothing like the copious merchant archives,
company records, toll registers and lists published by brokers of
pricey current in the bourses - the bread and butter of European
trade history - exists from Roman antiquity. The ancient histo-
rian hasto make do with much less andmust bewilling andable
to combine the testimony of very different kinds of sources. He
or she cannot afford to ignore the evidence of any one group
of material. Papyri, ostraca, wax tablets, Roman law, the classi-
cal literature, inscriptions and archaeological evidence must all
beconsulted andmadetocontribute tothefinalimage. Nosingle
group ofsources cansettheanalysis ona firmfooting^ Inscriptions
mentioning merchants, for instance, provide very little detailed
information andarein anycaserelatively fewinnumber, leaving
large tracts of the Empire completely in the dark. Archaeology
hasproduced abundantevidenceoftransport ofgoodsaroundthe
Mediterranean, but mainly through finds ofpotsherds in field sur-
veys and excavations. This tells us of transport of oil and wine
in clay containers and trade in pottery. Perishable containers and
goods are, on the other hand, often invisible. Thehugenumber of
grain ships bringing supplies from Egypt andAfrica to Rome have
lefthardlyanytraceinthearchaeological record,whereastheprob-
ably smallernumberofshipssailingwithcargoesofoil ancTwine
dominate thewreckevidence. 3 Cloth, animportant commodity in
pre-industrialtrade,isalsolargelyunaccountedforbyarchaeology.
Though clearly of great worth, the distribution patterns reveafed
3 Cf. Parker 1992.
2
PROLEGOMENA
by archaeology cannot easily substitute for our missing trade
statistics.4
The general inadequacies of the evidence accentuate the role of
conceptualisation in historical research. Too often historians for-
get that the sources are not identical with history; they are only
the fragmentary remains, the traces of the past world which we
seek to re-create in our writings. Sources are, in other words, not
self-explanatory. They must be interpreted in order to bring us to
the ancient reality. This has, in fact, always been acknowledged by
most practitioners ofthe historical profession. But it wouldbe fair
to say that, with some very honourable exceptions, the discipline
has neglected the development of interpretative tools. The way
that history was institutionalised as a discipline across universi-
ties in Europe and America in the nineteenth and early twentieth
centuries focused on promoting the detailed examination and crit-
ical scrutiny of sources. How to combine the observations into a
coherent whole was, on the other hand, seen as a function of the
personal inclinations of the individual scholar. Romanticist histo-
rians claimed to be artists, working on the basis of their personal
intuition and "divine" inspiration.5 Scholars of a more positivist
bent agreed. They tended to regard the "art" of interpreting the
evidence as a subjective activity, not really part of the proper "sci-
entific" pursuit of history, which should simply stick to the sources.
The latter was no solution; ultimately it was simply a way of giving
up on writing history.6
The historian of Hellenism, Droysen, followed by Max Weber
in the next generation, offered an alternative to this submission
to subjectivism. Intuition might be of service to historians dealing
withproblems andoccurrencesofwhichthey hadimmediateexpe-
rience. However, as soon as they had to deal with questions beyond
their personal horizon, their experience stopped being of any use
to them. Indeed it might often lead them to make false judge-
ments and introduce anachronistic features to their analysis - the
4 Whittaker 1989. 5 Meyer 1910.
As the "founder"of the modem historical profession in Denmark, KristianErslev quite
possibly did, cf. Tandrup 1979. A small booklet (Erslev 1911),makinga sharpdistinction
between the critical scrutiny of sources and the subjective art of interpretation, has
continuedto provoke intensedebateamong Danishhistoriansright up to the present day.
See Finley 1985b, chapter 4 for a general treatment of the question.
 PROLEGOMENA
cardinal sin ofthe professional historian. Suchconcerns arepar-
ticularly pertinent with regard to phenomena either far removed
in time and place from the historian, or such as cannot readily be
seen from the observation of individual examples but only appear
from the statistical study of large masses of evidence. Both con-
ditions apply in the study of ancient economies. The response of
Droysen and Weber was to pay close attention to the formation
of concepts - our analytical tools - or theory. By broadening and
refining his conceptual and theoretical repertoire, the historian
acquires a method for improving andmaking explicit the assump-
tions that he brings to bear on the source material. Theory and
conceptual analysis are a means by which the historian controls
and compensates for his "deficient" experience.7
One example must suffice for demonstration. In common par-
lancemarket trade isalwaysthesame. Careful theoretical examina-
tion ofmarkets, however, reveals thatthey may operate in very dif-
ferent ways according to circumstances. The bazaar, for instance,
possesses a number of characteristics, which produce a distinct
pattern oftradevery different from thatofmodern, capitalist mar-
kets, as we shall see more fully explained in Chapter 3. For the
moment, a basic definition in ideal-typical terms will do. As a
form oftrade,thebazaarhastakenshapefromthemanyirregular-
ities characteristic of pre-industrial peasant societies. Compared
to modem markets, the bazaar isdistinguished by high uncertainty
ofinformation andrelative unpredictability ofsupply anddemand.
This makes the prices ofcommodities in the bazaarfairly volatile.
Asa consequence, theintegration ofmarkets isoftenlowandfrag-
He; it is simply difficult for traders to obtain sufficiently reliable
and stable information on which effectively to respond to devel-
opments in other markets. Considerable fragmentation of markets
prevails. From a modern perspective where trade has focused on
overcoming these obstacles to commercial integration by creating
a more transparent and predictable environment, this condition
has often been mistaken for a seeming chaos and lack of organ-
isation. Most merchants operating in a bazaar environment'are
Droysen1977,esp.217-219and285-393;Weber1922:146-290;Finleyiq8'ib.chapter
4; Veyne )976b.
PROLEGOMENA
small and household-based. However, they were and are not pas-
sive victims of uncertainty. Instead they tend to pursue strategies
which do not aim to eradicate the general conditions of uncertainty,
but rather reproduce them as part of the solution. One such strat-
egy, as observed by Clifford Geertz, is clientelisation. Merchants
in the bazaar seek to cultivate personal and lasting relations of
exchange with particular business partners. This serves to shelter
them against the all-pervasive risks and uncertainties of the bazaar,
but on the basis of specially trusted and favoured relations. The
marketasa generalisedsphereofexchangeis not available,or only
in a very limited way, to the bazaar merchant. His is not a level
playing field. Better opportunies might, for instance, be available
in a different market. However, if the merchant lacks his network
of personal alliances, he may not be able to make use of those
opportunities and will therefore prefer to remain in the market
he knows. Fragmentation of the market is not simply an imper-
fection; quite the reverse, it is reproduced by the way the bazaar
functions.
Describedin this way,thebazaarasa model ofmarketsis bound
to be of particular interest to the student of pre-industrial trade. It
provides an attractive formula by which to treat the many irregu-
larities characteristic of a world based on slow-moving animal and
weather-boundsea transport. Yet, even here is it possible to make
distinctions.Inthisstudy,thebazaarisusedasa conceptto describe
markets in complex agrarian societies before the development of
(early) modem capitalism with its links to mercantilist privileges
and later liberal programmes of laissez-faire economics. These
policies helped to strengthen the position of the middleman, often
a man of dubious reputation and status in aristocratically domi-
natecl societies, and saw a development towards more stable and
integrated markets during the seventeenth to nineteenth centuries
within the European-led,for lack of a better word, worlcl-system.9
Theoretical and conceptual reflection enables the historian to
cultivate his analytical imagination and increases his capacity
8 Geertz 1979is thebesttheoreticaldescriptionofthe bazaar.
9 Cf. Wallerstein 197411. Steensgaard 19733 and Persson 1999 on the link between Euro-
pean commercial policies, institutional innovations and the gradual improvement in the
integration of markets; Bang 2006, for further discussion.
 PROLEGOMENA
for perceiving historical differences. 10 Traditionally historians
have described Greco-Roman antiquity with concepts taken from
the "Western" historical experience. For a long period clas-
sics was seen as the epitome of civilisation and constituted the
backbone of elite education in Europe and America. As this posi-
tion was lost in the cluster of processes normally referred to as
modernisation, a gap began to open up in people's minds between
the European present and the Greco-Roman past. Increasingly
Greco-Roman culture has come to be seen as a foreign place and
theexperienceofmodernity asa correspondingly inadequateguide
to the ancient past. This has set classicists and ancient historians
freetoexplore thelessfamiliaraspectsoftheancientworld. They
have in particular sought inspiration from anthropology. This dis-
cipline, having developed a rich theoretical tradition, has proven
a virtual gold-mine for ancient historians in search of organis-
ing concepts. Indeed, the anthropological work of Clifford Geertz
greatly informs the use in this study of the bazaar as an ideal-
type. But the relevance of anthropological research for ancient
history also has very real limits. Anthropologists have tradition-
ally eitherfocused on so-called primitive societies or studied tradi-
tional, most commonly peasant, communities in isolation from the
outside world. " This is a problem for the student of an urbanised
high culture, which depended on its ability to break the isolation of
theindividualpeasantvillagesandclaima partofagricultural pro-
duction for theconsumption ofurban dwellers andspecialists. The
isolated units of anthropology are too simple for the ancient real-
ity. '2 Economics might offer another possible theoretical avenue
to the student of Roman trade. In particular, development eco-
nomics, dealing with societies where agriculture still holds a very
prominent position, has something to offer. But much has already
changed in the third world. It doesnotrepresent a pristine version
10 Bloch 1967:44-8i;Veyne 1971;Finley 19855.
" A local emphasis, failing to take proper account of intercity/long-distance trade, is a
particular weakness of Geertz' early work (1963) on the bazaar.'The 1979 treatment
is moreawareoflinksbetweenmarkets,buttheattentionis still pointed'towardslocal
links.
Finley 1975:Chapter6 HardenandPurcell 2000strugglewiththisproblem; theyhave
difficulties combining their micro-analytical approach with a clear understanding of
wider processes such as state formation; see Bang 2004.
PROLEGOMENA
of pre-industrial society. In addition, economics as a discipline is
more orientated towards the future than the past. Its objectives are
different from those of history.
Both anthropology and economics offer valuable insights to the
ancient economic historian. This study draws on some of these.
But neither of the two disciplines has focused its attention on the
problems most relevant to the student of the ancient world - a
pre-industrialhigh culture. The ancienthistorian is thrust back on
himself or herself to develop an adequate theoretical framework
more thoroughly grounded in a historical experience relevant to
his/her own subject. This brings us back to Weber's project. His
ambition was to ground the social sciences securely in historical
experience. Theories developed in relation to the modem world
had to be complemented by concepts developed to deal with pre-
industrial conditions - a historical sociology. Weber's tool was
comparativehistory. 13By studyingothersocietieswhichresemble
the Roman world more closely, ancient historians acquire a more
secure basis for their own interpretations. Comparative history
provides them with the historical experience they lack.
The number of relevant historical parallels to the Roman Empire
is relatively limited. Pre-industrial history does not present us with
many giant, multiethnic agrarian empires sufficiently well known
to warrantcloser comparison. The last point bears emphasis. Pre-
Colombian empires in the Americas, such as the Inca and Aztec,
though smaller, might for instance be considered potential can-
didates for comparison. But our knowledge of these polities is
considerably less than of Rome. The sources are fewer and schol-
arship much less copious. The same goes for a number of ancient
and medieval imperial formations. 14 Anyhow, this is not to say that
these are not interesting cases. 15 But from a pragmatic perspec-
tive, the best candidates for comparison would seem to be found
in the context of sixteenth- to eighteenth-century Asia. Ottoman,
Ming/Ch'ing and Mughal history all offer attractive comparisons.
13 Weber 1972. See Skocpol 1979; Skocpol and Sommers 1980 on the analytical qualities
of comparative history.
14 The volume edited by Alcock, D'Altroy, Morrison and Sinopoli (2001), while interest-
ing, did reveal the huge discrepanciesin the densityof information betweenthe Roman
Empire and many potential comparisons.
15 See Scheidel 2008 for an interesting attempt to compare Rome and Han China.
7
 PROLEGOMENA
These empires have size, dense documentary record and extensive
historiography speaking in their favour.
But the choice is not merely a matter of expediency. The early
modern timeframe means that all three imperial societies were,
in turn, confronted withexpansive,Europeancapitalism.Thishas
causedMughal, Ottoman andChinesehistoriansto debateto what
extent trade and markets in these empires resembled or differed
from European (see Chapter i). The challenge facing these histori-
ographies has in many ways been the same as that confronting the
Roman historian: to find a way to account for vibrant commercial
networks and considerable exchange without denying the exis-
tence of significant differences to the emerging forms of modern
capitalism, in other words to avoid the Scylla of primitivism and
the Charybdisofmodemism. 16The historianofRoman trade can
surely leam from these efforts. This study draws on the experience
from all three empires, but Mughal Indiahasbeen singled out for
more detailedcomparison. Thischoiceisnotwholly arbitrary. The
character and shape of the historiography of the Mughal Empire
is such as to be particularly helpful in suggesting solutions to'the
problems which have occupied historians of the Roman economy
and trade - as will appear from the following chapters. Suffice it
here to note the availability of an extensive literature on trade in
the IndianOceanandtheprominent position claimed by a model,
positing a close link between state taxation and commercialisa-
tion inexplaining economic developments in theMughalEmpire,
much asin Hopkins' influential interpretation ofRoman economic
history.
A Roman-Mughal comparison may seem to militate against
the instincts of traditional historical research. It disregards' both
the unity of time and of space. Trans-epochal as well as cross-
cultural comparisons are often viewed with suspicion. '7 This is in
many cases unwarranted. Neither cultures nor epochs are "sacred"
boundary markers. They are analytical, not arbitrary, I hasten to
add constructs, useful generalisations designedto throw light on
specific aspects of reality. Such categories must be designed and
Bangetal. 1999;SeeBangandBayly2003forsomepreliminarycomparisons.
17 Even Tilly 1984. Bloch 1967: 44-81 is more nuanced.
PROLEGOMENA
adjusted to serve the historical inquiry. From an economic point
of view, imperial capitals like Rome and Agra - giant cities of
the pre-industrial world with population figures reaching into the
hundreds of thousands and financed by an enormous expenditure
of imperial tribute - clearly have much more in common than
each of them has with a contemporary smallish peasant village
within their respective cultural orbits in, say, Anatolia or Bengal.
"L'unite de lieu n'est que desordre. Seule 1'unite de probleme
fait centre", as Marc Bloch once remarked. '8 This approach does
not exclude recognising differences. The production of calicoes in
Bengal during the Mughal era represents one of the most extensive
pre-industrial examples of cloth production. A significant, though
far from the biggest, part ofproduction went to foreign exports. '9
Roman products such as wine and glass did find their way out of
the Empire, but the Roman economy was without a similarly large
export trade. In thatrespect, India,in spite ofhavingoften enjoyed
a reputation for being the quintessential traditional society, would
seem to be not less commercialised than the Roman world, rather
the opposite. This makes Mughal India a useful upper benchmark
for discussions of Roman trade and markets. Categories, in effect,
must be handled pragmatically and tailored to fit our analytical
needs.
This takes us to the problem of trespassing. Only rarely can
comparative historians aspire to be experts in more than one field.
Therefore they must depend, to a large extent, on the work of
colleagues in other areas of research. Inevitably, they are bound
to make some mistakes, just as their understanding of parts of
the "foreign" material undoubtedly will leave much to be desired.
None of this should be seen as vitiating the endeavour. It bears
emphasis that the aim is not to produce a comprehensive and
detailed account of Mughal India. Not every detail or nuance is
important to the comparative historian. Emphasis must be placed
on the aspects which can help to elucidate the problems in which
he/she is interested. The same goes for scholarly disagreements.
No historiography is ever fixed. New generations tend to produce
new, competing versions of the past. Any historical question of
18 Bloch 1934: 81.McNeill 1986:35 and84;Aymard 1990. '9 Prakash 1998.
 PROLEGOMENA
morethanmoderateinterestis boundtobecontested.Inpracticeit
is not possible, nor is it desirable,to accountfor every conflicting
view in the interpretation of the comparativeexample. In general,
I have chosen what seemed to be the better-argued case, even if
controversial, and restricted myself to indicating only some of the
major fault lines in scholarly debateson the Mughal Empire. The
aim, it should be remembered, has beento use Mughal history to
develop concepts for Roman history. 20
Remainingtobeexplainedarea fewmattersofterminologyand
chronology.Theterm "tributary"empiremayrequireclarification.
TributumdesignatesinLatinthetaxesowedbytheprovincestothe
Roman imperial state. Duringthe 1970$andearly 1980$the term
gainedcurrency amongMarxist scholars to describethe predom-
inance of the political power to tax in mobilising the agricultural
surplus ofpeasant producers. This distinguishes a tributary empire
fromthemodemcolonialempiresoftheeighteenthandnineteenth
centuries. The latter treated their dependencies much more with a
view to commercial exploitation. Provinces provided the Romans
withtaxrevenue;coloniesservedmoreassuppliersofrawproducts
and consumers ofmetropolitan exports. 2' Chris Wickham gave the
concept a well-directed twist, when he suggested it be used to dis-
tinguishcentralised, imperial surplus extraction from rent claimed
by local, decentralised and independent feudal lords. 22 It is in this
sense that the term has been adopted in the present study, though
in an undogmatic way and without subscribing to the broader
Marxistframework. In general"tributary"seemspreferableto the
empty and less precise phrase "traditional" empire. Occasionally
'tributary"hasbeenusedinterchangeablywith"agrarian"or"uni-
versal" empire. The term "universal" may help to clarify further
the notion of empire used here. It is generally reserved for vast
extensive hegemonic polities that manage to dominate and absorb
most states within their orbits. This sets a tributary empire apart
from the early modem absolutist monarchies which developed on
the basis of European feudalism and formed part of a pluralistic
20 Paludan 1995: 197-203 is the best available discussion of this problem. See further
Goldstone 1991: xxvii and 50-60.
21 SeeHobsbawm1987,chapter3 fora treatmentofcolonialempires.
Wickham 1985;Wolf 1982;Amin 1973.
10
PROLEGOMENA
state system maintained not by a clear hegemon, but by a balance
of power. 23
Another terminological "mannerism" employed here is the con-
cept "agrarianatesociety". The term was coinedby the innovative
comparative historian of Islam, Marshall Hodgson, who used it
to emphasise that literary high culture and cities co-existed with
peasant agriculture in the pre-industrial world. It was not simply
an "agrarian" society. The next issue is chronology. The emphasis
in this study is on the Roman Principate. However, occasionally
discussion of examples and evidence from both the late Republic
and late antiquity has been included in order to throw light on the
problems under examination. Though clearly seeing changes, the
broader timespan was also marked by one fundamental continuity:
the presence of a vast tributary empire in the Mediterranean world.
Finally, a word about citation of Greek and Latin texts. These
are normally reproduced in English translation as a help to the
reader. The argument, however, is based on the language of the
original. Sometimes, therefore, I have not indicated the source of
the English text, so as to be free to make the changes, modifications
and adjustments I deemed necessary, whenever the translation did
not originate with the present author. In most such cases the text
has been taken from the standard translations of the Loeb series.
The first part of this study seeks to identify and develop the
most suitable comparative framework for the analysis of trade in
the Roman Empire. Chapter I treats the debate on the Roman
economy as an expression of a set of problems common to pre-
industrial studies in general: to whatextent should the evolution of
European capitalism be treated as a general template for analysing
trade in other complex agrarian societies? The evolution of a sys-
tem of emergent nation states in early modem Europe is now seen
as instrumental in the development of modern capitalism. Fierce
competition and rivalry forced governments to promote economic
development and often trade in particular. The Roman imperial
government was under no similar constraint. The ancient process
of state formation had united the greater Mediterranean under the
rule of the Romans. Interstate competition was a much weaker
23 Tilly 1992; Watson 1992.
II
 PROLEGOMENA PROLEGOMENA

force. The Roman state was already the most powerful and had eft to the initiative of merchant communities and their capacity
much less need to privilege commercial groups; it could concen- :'or self-organisation. Chapter 3 analyses market integration and
trate on taxing its vast population, particularly the peasant masses. argues that Roman merchants would generally have been less able
In that respect, the Roman Empire resembles the early modem to integrate markets than their later early modern European col-
empires of Asia, the Ottomans, the Mughals and China of the leagues. Instead, it suggests the bazaars of the Mughal Empire as
Ming and Ch'ing dynasties. These empires were all tributary and a useful parallel, presenting an image of more fragmented inte-
the chapter concludes by suggesting them as constituting a more gration, many local monopolies and relatively greater uncertainty.
proper category within which to analyse the character of Roman Chapter 4 examines the question of uncertainty in greater detail
trade. by looking at protection and customs duties in the Roman and
Chapter 2 explores the character of economic integration in the Mu.ehal Empires. Though customs claimed by the state were pre-
Roman world. Through a comparison with surplus extraction in dominantly fiscally motivated andfrequently predatory, the ability
Mughal India, it is shown how the formation of a tributary empire of merchants to smuggle and to circumvent the harshestcollection
affects the main driving forces behind long-distance or intercity points checked abuse by officials and kept the burden of indirect
trade. The imposition of an imperial tax saw a concentration of taxes within bounds. In the absence of elaborate state guaran-
wealth in the hands of the imperial government and aristocratic tees, the merchants of the bazaar developed their own private,
elites. Political consumption of the agricultural surplus, however, social institutions to tackle the uncertainties of commercial life.
depended on the availability of markets and long-distance trade. Chapter5 examinesthese in terms of communal, religious organi-
Intensified surplus extraction therefore went hand in hand with sations and the role of the household in supplying trading capital.
increased commercialisation. Market formation was driven by The epilegomena conclude the study; they sum up the argument
politically determined consumption rather than economic com- and identify some themes and avenues for further research with
petition. This argument is supported by a hypothetical quantitative respect to the development of imperial styles of consumption.
model of the Roman economy.
The second part of the study, "Imperial Bazaar", moves on to
examinetheinstitutionalconditionsoftradeandtheshapeofmar-
kets produced by the tributary empire. Its basic aim has been to
redress the balance in our image of trade under Roman hege-
mony. While Roman historians have frequently been impressed
by the image of tranquil order projected by the Empire, Mughal
historians have, in comparison, paid much more attention to the
predatorydimensionsofimperialrule.Bothaspects,however,need
to be accommodated within our understanding. Increased surplus
extraction, the creation of imperial coinages and the imposition
of imperial law, all facilitated trade and markets. But there were
clear limits to this process. Many irregularities still continued to
characterise the world of trade under the imperial peace. State
personnel did not only offer protection, they were also a source
of predation and insecurity for merchants. The administrative and
judiciary machinery was of very limited extent. Much was still
12 13
 PART I

THE ROMAN EMPIRE AND THE COMPARATIVE

STUDY OF PRE-INDUSTRIAL SOCIETY

Ideally we should create a third discipline, the comparative study of literate

post-primitive (if I may), pre-industrial societies.
M. I, Finley, Anthropology and the Classics

After all, something has changed in the course of human history. We need terms
which can size the differences.
E. Gellner, Plough, Sword and Book
 BEYOND THE ANCIENT ECONOMY^
TRADE IN THE ROMAN EMPIRE AND THE
PROBLEM OF COMPARATIVE HISTORY

AberdieseAnalogienmitmittelalterlichenundmodernenErscheinungen,schein-
bar auf Schrift und Tritt vorhanden, sind zum nicht geringen Teile hochst
unverlasslich und oft direkt schadlich fur die unbefangene Erkenntnis. Denn
jeneAehnlichkeitenkonnenleichttrugerischeseinundsincles tatsachlichnicht
selten.
Max Weber, Agran'erhaltnisse im Altertum, p. 4
En Asie, on a toujours vu de grands empires; en Europe, ils n ont jamais pu
subsister.
Montesquieu, L'Espritdeslois, 3. 17.6

Rome - the battlefield of Orient and Occident

Some years ago the world of classical scholarship saw one of

those rare genuine sensations. A manuscript was found and pub-
lished containing a set of very full notes, taken by father and
son Hensel, of Mommsen's lectures on the history of the Roman
emperors. Here atlast wasthe almost mythical, long hopedfor and
neverpublishedfourthvolumeofMommsen'slegendaryRomische
Geschichte. After having dealt, in three volumes, with Rome from
the earliest times up to the battle of Thapsus in 46 BC, the history
hadbeenplannedto proceedwith a treatment of the political his-
tory of the Romanemperors. It never happened.The subjectheld
little appealfor Mommsen. In his viewthe systemofmonarchical
government had workedbest "whensecond-ratemen were at the
head". ' Instead, a gap was left in the series and Mommsen con-
tinued his history of Rome with volume 5, published in 1885,
which contained a ground-breaking survey of the conditions pre-
vailing in the provinces. Rather than the decadent and mediocre
emperors, long familiar from the narratives of Tacitus and

Mommsen 1992: 231 (my translation). Skydsgaard 1997.

17
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
Suetonius, the matter of real importance and interest, as Momm-
sen saw it, in the history of this period was the gradual spread
of Greco-Roman civilisation throughout the provinces of the
Empire. 2 Among thechapters on the eastern provinces he included
a treatment of the border zone between Rome and its formidable
neighbourand adversary,the ParthianEmpire. Whatclashedhere
were not just two ancient, competing empires but two funda-
mentally opposed principles of society and culture - the city-
based, Western-style bourgeois, Greco-Roman civilisation versus
the Eastern civilisation of the Parthians, almost without towns and
based on the despotic control of large masses of toiling peasants
living in villages. The border thus took on an almost mythologi-
cal dimension in Mommsen's analysis, constituting a demarcation
line between East and West, Orient and Occident.
A few years earlier that great English historian of compara-
tive jurisprudence, Sir Henry Sumner Maine (1822-1888), had
defined the difference between East and West as a question of
societies basedon status rather than on contract - hierarchy versus
liberal egalitarianism. As the ancien regime gradually gave way
in Europe during the eighteenth and nineteenth centuries and the
continent established world dominance, the civilisations of Asia
came to be seen as specimens of an unchanging traditional and
hierarchic social order, which history somehow had left behind.
Hegel, and later Marx, placed the Orient outside the dialectical
progression of world history. China and India were seen as static
social systems betraying no signs ofdevelopment, always remain-
ing the same. They were without history.3 Thus, to Maine as to
many others, Indian society with its age-old Vedic cultural tradi-
tion, composedin an Indo-Europeanlanguage,offeredthestudent
of history a glimpse into the long bygone past of Europe. What
he saw was a society consisting of quasi-autarkic villages made
up ofpatriarchal households with the division of labour organised
along caste lines and only limited influence exerted by the market.
He believed that it was possible to detect the same traits in the
.
' Mommsen 1904, Introduction.
Hegel 1996: 122-123; Marx's views on Indiaarebestexpressed in hisdispatches to the
New York Daily Tribune, e. g. 10 June 1853 and 22 July 1853, easily downloaded from
the Web or available in e. g. Marx and Engels 1960.
l8
BEYOND THE ANCIENT ECONOMY?
earliest remnants of Roman law. But Roman law had not remained
ciuiulit in the fetters of status society. It had continued to progress
iiii. l i I it achieved a fully developed system of rational law based on
contractual obligations and rights. Thus Roman law represented a
brkl.^e, figuratively speaking,betweenthe old status-baseddevel-
opmental stage and modern Western rational, commercial society.
By studying the evolution of Roman law and its culmination in the
classical Romanjurists, it was possible to follow the birth process
of the founding principles of Western civilisation.4
So Greco-Roman civilisation, or rather its greatest achieve-
irsents, were in many waysperceived aspresenting a developmental
history similar to that of Europe since the Middle Ages. This is,
of course, a way of thinking which ultimately traces its roots back
to Renaissance ideology, proclaiming the new times as a rebirth
of Greco-Roman antiquity, and which later received a new boost
from the art historian and archaeologistJohannJoachimWinckel-
mann (1717-1768), who inspired the neo-classicist adoration of
the Greeks and to a lesser extent the Romans. Even in such an inclu-
sive, overtly anti-classicist, history of antiquity as that of Eduard
Meyer (1855-1930), whichalso insistedon the importanceof the
various Oriental or Levantine cultures as background and prede-
cessors of the Greek and Roman world, only this last civilisation
is depicted as progressing through all or most stages of historical
development. In contrast, Pharaonic Egypt is seen as history gone
wrongaftera promisingstart, a caseofOrientalstagnationcaused
by religious superstition and an oppressive clergy. 5 Thus the eco-
nomic historian, Karl Biicher (1848-1930), could hardly avoid
appearing strongly provocative when he dared in his Die Entste-
hung der Volkswirtschaftof 1893 to view Greco-Roman antiquity
in its entirety under the same heading, that of household economy
(Oikenwirtschqft), in order to distinguish it from the Middle Ages
characterised by a city economy (Stadtwirtschaft) and the mod-
ern period with its national economies (Volkswirtschaft). Meyer
immediately felt the need to rebut what he saw as a dangerous,
simplistic unilinear view of history, and chose an official meeting
4 Maine 1871, chapter i and 1861. Meyer1910:192-193, 247-252 and 1924:89.
6 Bucher1911:85-150.
i9
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
of German historians in 1895 to dismiss rudely Bucher's views
as those of an amateur, unfamiliar with the sources. Bucher's
description might, albeit with difficulties, fit the earliest, most
primitive, periods of Greco-Roman culture; but for later times
with the progress of civilisation, the extent of trade had become
much too large. Here one hadto employ his later economic types of
Stadt- and Volkswirtschaft. This analysis Meyer concluded with
the now legendary dictum that the ancient economy "in every
respect, cannot be considered modem enough. Only one should
avoid drawing on the nineteenth century for comparison, but use
rather the seventeenth and eighteenthcenturies."7
One of the classic debates of historiography had been born, the
struggle between "primitivists" and "modernists", which still, a
century later, continuesto hauntscholarlydiscussions,often under
the revealing name of minimalists versus maximalists, signifying
that the problem has often wrongly been reducedto one of quan-
titles, mainly of trade. Yet, Max Weber (1864-1920) in his 1909
articleAgrarverhaltnisseimAltertum,hadalreadyobjectedthatthe
main point ofBucher's models wasnot to assign antiquity to a first,
primitive stage of economy or to deny it the existence of historical
development. 8 On the contrary, Bucher's message was strongly
historicist and had developed from a critique of the English classi-
cal economists' attempt to explain all economic phenomena within
one comprehensive or universal theory. Hence he tried to identify
the organising principles, the economic structure, particular to each
historical society and epoch, rather than subsuming them all under
one unifying logic. The ideal-type of the oikos economy, in fact,
far from being a first primitive stage of economic development,
was rather an attempt to grasp what set antiquity apart as a his-
torical epoch among the group of advanced civilised societies.9
7 Meyer 1924: 141: "in jeder Hinsicht nicht modern genug gedacht werden kann. Nur
darf man nicht das neunzehnte Jahrhundert zum Vergleich heranziehen, sondem das
siebzehnteund achtzehnte"(my Englishtranslation). Seefurtherpp. 81-82 and 88-89;
Schneider 1990.
Weber 19240:7-13.
9 Bucher 1911. pp. 85-92 are instructive for his historicist critique of classical eco-
nomics. Meyer missed both this dimension of Biicher's thought and its separation of
oikos economies from the "primitive" economies of anthropology. In Biicher's scheme
the latter preceded the former category (pp. 92-116). Finley 1977: 315-317 saw this
clearly.
20
BEYOND THE ANCIENT ECONOMY7
Here it occurred to Weber that in certain respects the fundamen-
tal principles governing the economy of the Greco-Roman world
had more in common with the Levantine or Oriental societies of
antiquity than with the kind of economic formation one found
in early modem Europe. Therefore he treated the various ancient
civilisations known to him within a broad, common analytical
framework, trying not to sacrifice historical development or par-
ticulars by developing a range of more specific models to describe
the characteristics peculiar to each society and period. 10
Meyer, however, was blind to this historical aspect of Biicher's
and Weber's economic types and the accompanying compara-
tive perspectives, even though he himself insisted that history
was a discipline of the singular and the particular. In his inter-
pretation, Greco-Roman civilisation remained a quintessentially
Western bulwark against the Orient. In fact, the decline of the
Roman Empire was portrayed as a reversal, beginning in the East,
to more Oriental forms of society dominated by religion and
stiffening tradition. " This idea received its fullest exposition in
M. I. Rostovtzeff(1870-1952)andhismonumentalThe Socialand
Economic History of the Roman Empire of 1926. Here he set out to
supply the empirical underpinning of Meyer's basic interpretative
framework for Roman economic history during the Empire, not
least through an innovative attempt to bring the growing archae-
ological record to bear on the historical problems. To all intents
and purposes the history of the Roman Empire came to an end
with the beginning of the fourth century, from Rostovtzeff's point
of view. To him the rest was, indeed, silence, and the remaining
centuries, today known as late antiquity, were cursorily dealt with
and hurried to their tragic end under the heading "The Oriental
Despotism andthe DecayofAncient Civilization".Greco-Roman
civilisation in all its constituent forms had now ceased to exist. It
had become a foreign world - an Oriental world.
10 Weber l924a, Einleitung. See Bruhns 1985; Capagrossi Colognesi 1990; Love 1991;
Capagrossi Colognes; 1995; Bruhns and Nippel 2000; Nafissi 2005 on the Weberian
approach to ancient economic and comparative history.
" Meyer1924:142-160. 1910: l23-l53and 192-193 makes the connection of the Orient,
stagnation, tradition and religion clearer.
12 Rostovtzeff 1957, chapter 12. Most of his overall interpretative schema can be found in
Meyer 1924:142-160. This ought to temper the current ritual praise of Rostovtzeff's
21
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
Today such an evaluation seems somewhat dated, given the
flourishing state of late antique studies. One might even think in
termsofa paradigmshiftin classicsandhistoryhavingtakenplace
in the last half century or so. A seminal work in this process of
changewasA. H. M. Jones' study ofThe LaterRomanEmpire of
1964, which demonstrated that late antiquity could be studied as
an epoch in its own right and notjust as an era of decay and disinte-
gration or a prologue to the final liquidationofthe classical world.
Moreover,byreflectingupontherelativeweightofagricultureand
trade in the economy of the Empire, Jones came to hold a notion of
the ancient economy fundamentally different from that of Eduard
Meyer or Rostovtzeff. He characterised the Roman world, in late
antiquity as in other periods, as overwhelmingly agrarian. Trade,
so much emphasisedby the modernists, he relegated to a minor
position in the overall structure of the economy. Thus changes in
the patterns and volume of trade might not, after all, be indicators
of fundamental economic change. '3
In many ways the taskof teasingout the implications ofJones'
decisive shift of emphasis for our understanding of the econ-
omy of the Roman Empire remains at the very centre of schol-
arly activity and controversy. One result has been that people are
now less inclined to take a view of late antiquity as fundamen-
tally different from and completely alien to the preceding peri-
ods, and tend instead to emphasise many continuities, without
denying changes. '4 A good example, and one of the most per-
sistent attempts to come to terms with the agrarian "nature" of
the Greco-Roman world is, of course, the ground-breaking The
Ancient Economy of 1973, composed by Moses Finley, Jones'
successor as professor in Cambridge. Here developments in late
antiquity, such as increasedstate control, contraction in the market
sector, tying of peasants to the land and a crisis in the economy
of the towns, are seen as the result of limitations, imposed by the
utilisation of the archaeological material as against Finley's alleged neglect (Harris
1993: 14-15 and n. 26; Harden and Purcell 2000: 6l2). It is doubtful whether his basic
interpretation was in any determining way formed by archaeology. See Salmeri 1998
for a sobering reading of Rostovtzeff.
3 SeeJones [964,chapter2[, andmorebroadly 1974;Christiansen 1995:51.
14 Pleket 1990: 29-30; De Ligt 1993: 4; Garnsey and Whittaker 1998; Gamsey and
Humfress 2001.
22
BEYOND THE ANCIENT ECONOMY?
core institutions of Roman civilisation, in the Empire's ability to
respond to increased military pressure along the frontiers. This
is Finley, doing the sums: "The Ancient World was hastened to
its end by its social and political structure, its deeply embedded
and institutionalized value system, and, underpinning the whole,
the organizationand exploitation of its productive forces.'"5 The
Empire of late antiquity had now been firmly located within the
defining cultural characteristics of the classical world; during
the same exercise Finley had radically shifted the focus of our
understanding of the Greeks and Romans, stressing the "otherness"
oftheancientworldandrejectingtheallegedcloseaffinitybetween
the social and economic forms of antiquity and early modem
Europe.
Instead, Finley sought his inspiration from the, formerly unac-
ceptable, analysesof Bucher, Weberandtheirpupils, andin addi-
tion from the writings of various anthropologists. The economist
and anthropologist Karl Polanyi (1886-1964), whose seminar at
Columbia during the late 1940s and early 1950$ Finley attended,
wascited for the fundamental inapplicabilityof'market-centred,
modern economic theory to the study of antiquity; further,
Polanyi's notion of economics as the study of "the satisfaction
of material wants"echoesthrough Finley's discussionofthe issue
of the presence or absence of an economic policy, modern style, in
Greek and Roman states. 16 But there were limits to the usefulness
ofPolanyi's conceptualisation for Finley's project. Reciprocal gift-
giving and authoritarian redistribution, the two modes ofeconomic
organisation seen as characteristic of pre-modern economies by
Polanyi, were inadequate to explain the ancient reality. Even if the
Greco-Roman world did not operate on capitalist lines, markets
still played too much of a role in the economy to describe it in
terms ofPolanyi'salternatives. '7 Forthe concrete analyses,Finley
15 Finley i985a: 176. I6 Finley 19853: 26 and 160-161.
17 Cf. the important role ascribed to Athens' function asa commercial centre for the Aegean
world in generating income to the city after the loss of the empire (Finley l9»5a: 131-
134)- It is commonly forgotten that Finley rejected Polanyi's concrete models for not
allowing enough influence to markets and trade (Finley 1970). It is, therefore, not a
little ironic when Nafissi 2005, chapter 10 describes The Ancient Economy as a decline
into Polanyian orthodoxy and at the same time takes it to task for excluding the ancient
Middle East from the analysis. Yet, the latter was not included on the express grounds
23
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
turnedto inspirationtakenfrommore complexpre-industrialsoci-
eties. The growing body of peasant studies, with its emphasis on
locating markets in "traditional" settings, was a key influence. 18
In anotherplace, the workofthe anthropologistLouisDumonton
the Indiancaste system, Homo Hierarchicus,is utilised to embark
on a long discussion emphasising the role of hierarchy and social
status in Greco-Romansociety. 19Thus, newandexcitingcompar-
ative perspectives were opened up; and aspects or themes, often
associated with the exotic societies of anthropological discourse
or pre-modern Asian civilisations, were now allowed to appear
and come to the forefront in the picture painted of the classical
world.
Of course, Finley was far from alone in this. It is easy to for-
get this today, but the i96os and 1970s were also the era of de-
colonisation andthetiersmondiste movement, firedbythewritings
of Jean-Paul Sartre and Franz Fanon. New prominence was given
to notions ofcultural relativism, andthe limitations andhistoricity
of Western concepts were explored and highlighted by thinkers
suchasFoucaultand,in thefieldofeconomics,by thesundrycrit-
ics ofdevelopment theory. It would be difficult to maintain that all
that was said and done during these years was for the good. After
all, this was also the period that produced the various kinds ofvery
dogmatic Marxism and the horrifying crimes of a Pol Pot, trying to
realise the thirdway in a version profoundly different from current
thatmarkets seemed to play lessofa role andthat it resembled more a Polanyian scenario
(28-29). This view of the ancientNearEast may be mistaken (cf. n. 19), but hardly
makes the picture rendered of Greco-Roman antiquity any more Polanyian. Above ail
The Ancient Economy wasa Weberianstudy.
18 SeeFinley 19850,chapter4,n.2 (referringtoa nowclassiccollection ofpeasantstudies
edited by T. Shanin), further nn. 29, 33, 34, 70. In his essay "Anthropology and the
Classics"(1975), Finleyinsistedon a comparativecategoryofcomplexpre-industrial
societiesas the most relevant for classical studies ratherthan the primitive tribes often
the object of anthropological research.
19 Finley l985a: 43-44, thoughthere are remnants ofthe old Orient-Occidentdivide on
pp. 27-29; but there the question is of delineating the boundaries of the subject, not of
seeking out comparanda. Manning andMorris 2005 convincingly argue that thevarious
NearEasternsocietiescouldusefullybeincludedinourconceptoftheancienteconomy;
they were not asdifferent as was previously thought. But for Pinley (198511: 30-3 l with
n. 31), following Braudel, not ignoring him as Manning and Morris 2005: 15-17 claim,
the Greco-Roman world was a Mediterranean world, not a river valley world.
24
.?
BEYOND THE ANCIENT ECONOMY'!
i)!ilical usage. 20 Still, one should not lose sight of the liberating
erspectives - for which the close relationship between history and
anthropology to this day can work as a suitable symbol - that have
allowed new questions to come to the fore and not least sought to
v ide more historical and often also more theoretically refined
answers by exploring the limits of the experience and concepts of
the modern world. 2'
Hven if we stay within the field of Roman history, a long series
of studies, influenced by anthropology and comparative history,
springs to mind. I shall restrict myself to mentioningjust a few.
Paul Veyne asserted the primary importance of gift-giving, tradi-
tioiuilly the preserve of anthropologists, to the Greek and Roman
world in his Le Pain et Ie cirque of 1976. A few years previously,
Ramsay McMullen in Roman Social Relations (1974) had called
ulicntion to the high levels of extra-legal violence and the limited
e\!ension of official authority. Before that, Peter Garnsey had doc-
iimcnted the profound importance of status, social and legal, in
(lie heyday of classical Roman law, where Maine had interpreted
the elaboration ofthejuristic literature as a transition to a modem,
Ci.'alitarian contract-based form of society. Even more systematic
comparisons between the Roman and some of the more famous
Askin pre-modern empires and civilisations have been attempted.
Keith Hopkins drew continuously in his work on comparisons with,
especially, Chinese imperial history. 22 In his study of Patronage
iiiulcr the Early Empire (1982) Richard Saller included compar-
isonswiththe administrationofthe OttomanandChineseEmpires;
and Dick Whittaker (1994) has recently attempted to understand
the Roman imperial frontier through a model developed by Owen
Lattimore to explain Chinese history. Inspired by these studies,
a number of collaborative efforts in the field of Roman imperial
comparisons have also seen the light of day during the last few
years. 23
20 Finkielkraut 1988. -" Contra Engels 1990:131-142.
22 Hopkins l978a; 1980; 19833; and 1995-6.
23 Bang et al. 1999; Alcock, D'Altroy, Morrison and Sinopoli2001; BangandBayly 2003
and a project conducted by WalterScheidel(2008) comparing Rome and Han China.
25
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
Returning to Rostovtzeff?
All the same, these tendencies have hardly gone uncontested. For
instance, in a recent collection of essays, Inventing Ancient Cul-
ture, the editors set out to restore the prominenceofGreco-Roman
antiquity in our understanding ofthe birth ofmodernity, whichthey
findhascometobetoostronglyassociatedwiththeEnlightenment
era. 24 Certainly, in the field of social and economic history debate
hasbeenheated,to saytheleast, from thevery beginning,andone
cannot help but feel that it was somewhat premature when Keith
Hopkins,introducinga collection ofarticles presentedto Finleyin
1983,termedFinley'sandA. H. M.Jones'jointly developedinter-
pretation "the new orthodoxy". After all, that very collection did
not display such a consensus. The impression was rather that of a
multiplicity of viewpoints and strong disagreements. 25 Even if we
allow thatThe AncientEconomyhasoccupieda dominatingposi-
tion atthevery centre ofscholarlydebate,the directionofresearch
hasto a largeextent beendecidedby Martin Frederiksen'sdismis-
sivereviewin the JournalofRomanStudies.Colouredby a general
distrust of the theoretical reflections in which Finley so excelled
and by a strong positivist inclination, he set out to show that the
ancient evidence revealed a more complex and varied reality than
allowedfor by Finley's simplifyingmodels. 26
That sort of objection, however, is banal. No treatment of a
bygone world, however detailed, can ever claim to be more than
a simplification. The important question to ask is whether in the
chosenperspective a specific analysis seems to be a fair approx-
imation of the ancient reality for which it seeks to account. Seen
in this light much of the criticism has been less than success-
ful. 27 Nonetheless,Frederiksen'spositivistobjections havestruck
a strong chord in academe, and have, of late, received somewhat
24 Golden and Toohey 1997. Schiavone 2000 elegantly demonstrates the futility of such
anenterprise. 25 Garnsey,HopkinsandWhittaker1983.
26 Frederiksen 1975. SeeDrexhage, Konen andRuffing 2002: ii and21 for some very
recent echoes of Frederiksen's objections. Bang, Ikeguchi and Ziche 2006 for a collec-
tion of papers which, on the contrary, advocate the need for the conscious use oftheory
in the study of the ancient economy.
27 Erdkamp 2001; see Saller 2002 for some cautionary observations.
26
BEYOND THE ANCIENT ECONOMY?
unexpected assistance from the present postmodernist dislike of
generalised explanations. One view asserts "that whatever clas-
sificatory or analytical system we use has to accommodate, in a
way that the rnonocolore and essentially synchronic constructs of
Blicher and Finley cannot, a possible extreme diversity of types
of activity within the same society". 28 Another book, edited by
Helen Parkins, confidently declares that we are now Beyond the
Consumer City - the model Finley used to describe the economic
foundation of Greco-Roman urbanism- andemphasises the need
for a diversity of approaches. 29 One can readily applaud Parkins'
desire to have ancient urbanism examined from a variety of view-
points. After all, no model can ever hope to be exhaustive; its
application is of necessity restricted to a limited number of prob-
lems. But it is less than clear why an openness towards a plurality
of approaches would require the theory of the consumer city alone
to be banned or left behind. Certainly, Parkins presents no very
convincing case to back up her claim. On the contrary, in her own
contribution to the book she is hard pressed to find her way. "None
of this contradicts Weber's consumer city model", she has to admit;
indeed her analysis sets out from a perspective "that incidentally
underlines the validity of Weber's model while adding some sig-
nificant refinements". 30 One could reasonably have left it at that.
Even so, Parkins still insists that the level of involvement in the
urban economy of the land-based elite was somehow too large to
be accommodated by Finley and Weber's model. The logic is far
from compelling.3
There is something a little doctrinaire about this kind of reac-
tion, yet the style of argumentation is widespread. In their book
The Corrupting Sea,PeregrineHardenandNicholasPurcell man-
age within the space of two pages on the one hand to concede
that the understanding of the city as a centre for aristocratic con-
sumption of an agricultural surplus is whatbest sums up the phe-
nomenon of ancient urbanism and on the other to reject the model
as "dated": for there was allegedly more growth in the economy
than allowedfor by Finley.32 The issuesraisedby Parkins,Harden
28 Davies 1998:241. .>9 Parkins 19975,preface. SeefurtherParkins 1998: II-I3.
30 Parkins l997a, citations from pp. 107 and 84. 31 Pace Mattingly 19973: 215.
32 Harden and Purcell 2000: 106-107.
27
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
and Purcell are not unimportant. In the final analysis, however, the
model survives and this is, more or less, openly admitted. Thus
Harden and Purcell depend to a large extent on Morley's thought-
provoking study of classical Rome - the quintessential consumer
city - which argues that the immense buying power created by
the politically directed concentration of economic resources from
an extensive empire in the capital also acted as a stimulus on its
Italian hinterland, promoting more intensive forms of agricultural
production.33 In Parkins' case, an acceptance of even extensive
aristocratic investment in the urban economy does not alter the
fact that the wealth of that class was grounded in agriculture or
that the cities depended for the satisfaction of their consumption
needs on the surplus from its landed estates. A strong presence in
theurbaneconomywouldratherserveto putevenfurtheremphasis
on aristocratic dominance in the city and thus underline the cities'
generaldependenceonthe largehouseholdsoftheelite. 34Farfrom
having made the consumer-city model obsolete, the critics seem
to be exploring aspects or areas of activity where the model can
accommodate several variations. Not polarised confrontation, but
constructive engagement is what is neededhere. This would allow
a more fruitful debate about the model to emerge where attention
could be clearly concentrated on charting variations, identifying
sub-types and adding new dimensions to our understanding of
ancient urbanism, for instance by exploring the formation of urban
networks and hierarchies.35
Further examples of misdirected critique at Finley and the
"primitivist" postion need not detain us here. Enough has already
been said to make it clear that the unconstructive, sometimes
almost extreine reaction against "the Cambridge orthodoxy" has
been accompanied by a tendency to blur discussions and more
importantly to allow a much less precise use of analytical con-
cepts to reassert itself. The unfortunate outcome has been that the
debate on the ancient economy has become stuck on a well-wom
33 Morley 1996.
34 Parkins 19970: 88-89 is not far from that conclusion. Wallace-Hadrill 1991 and
Whittaker 1993, chapter 12 and 19946 are very clear aboutthis.
35 Whittaker 1995 is an interesting exploration of the limits of the explanatorypower of
the model.
28
BEYOND THE ANCIENT ECONOMY''
track as attempts are made to pinpoint its precise location on a lin-
ear spectrum between the two opposites of primitive and modem,
as indeed recommended by Frederiksen in his review. 36 In that
analytical game, less is never more. 37 A tendency has emerged
in current scholarship almost automatically to support taking a
more optimistic view of the ancient phenomena, be it rationality,
growth, trade, market integration or the economy in general. 38 No
wonder then, that Rostovtzeff is again celebrated as the great-
est economic historian of Mediterranean antiquity and attempts
have been made at resurrecting his general interpretation of eco-
nomic development in the first centuries AD. Oncemore the story
has been told of how a flourishingempire-widemarket developed
under the favourable influence of the Fax Romana, in which the
provincescompetedformarketsharesandgraduallyoustedItalian
products.39 More broadly,this return of"modernism"canbe seen
in what one of the veterans of the struggle has called "a drift back to
the bad old ways when huge trade, industry and capital formation
16 Frederiksen 1975: l7l. Whittaker 1995; Bang l997a; and Drinkwater 2001 warn against
this tendency.
37 Attempts at quantification, for instance, often incline towards the highest possible
estimate, e. g. van Minnen 1986 (extrapolation from very thin and fragmentary material
to postulate a very large global estimate of the Oxyrhynchite cloth trade, though we
would expect seasonal fluctuation rather than permanent high activity); Wilson 2001
(twenty-two household establishments, with small fulling installations in Timgad dated
broadly within a bandof three centuries, are taken as a sign of large-scale activity. Yet,
the twenty-two individually small establishments are unlikely all to have existed at the
same time, much change happens over three centuries, and even if they did, at five
members per household, the entire "industry" could be accounted for by some 100-
120 persons); Adams 2007: 227-228 (focusing on a Ptolemaic village customs register
indicating brisk activity in one month. But, as he also notes, this may be due to a post-
harvest peak in grain transports. The more numerous Roman customs registers indicate
less intense activity to be the norm, cf. Alston 1998 for a more cautious analysis).
3 Harris 1993; Greene 1994: 25-26; Harris 2000 and Temin 2001 for systematic surveys
of the economy and trade of the Roman Empire seeing more growth, trade, markets and
development than Finley allowed. Harden and Purcell 2000 simply define exchange
(market, gift, redistribution) as the central characteristic of pre-moclern Mediterranean
civilisation.
yl Tchernia 2006 for a critique of this trend. A new example of the continuing cache of
Rostovtzeff's general interpretive framework is provided by Rizzo's discussion of the
changingpatterns of imports to Rome (2003: 230) as reflected in ceramic assemblages,
seeing Italian production decline in the face of provincial competition. Carandini 1983^
1988; 1989 are the most outspoken examples. But elements can also be found in e. g.
Pater.son 1998; von Freyberg 1989, chapters 2 and 3. 5 and in the very interesting and
balanced attempt by Lo Cascio 1991 to "unite" Rostovtzeff and Weber. For Rostovtzeff
as the great historian, see note 12 above.
29
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
were discovered in every petty street-corner workshop of Roman
towns".40
Two examples will suffice. One deals with the production of
Italian terra sigillata, the mould-cast, red-slip ceramic tableware
found by archaeologists in widespread locations during the late
Republic and early Empire. The other is a broader discussion of
the market in the Roman world. Let us take terra sigillata first. Sur-
veyingrecentarchaeologicaldevelopments in ourpictureofItalian
production centres, Philip Kenrick concluded: "The level oforga-
nization and the closeness of relationship which this implies seems
to me to justify us in regarding this as an industry in an entirely
modern sense. "41 Yet, one looks in vain for any attempt at a close
definition of modem industry that would justify the label in this
particular context, especially in view of a whole series of studies
that have emphasised the large differences to the modem world
in the production of ancient fine ware. 42 Apparently the mere
existence of branch workshops, maintaining connections with a
mother unit, is taken as sufficient "proof. But we find Assyrian
merchants establishing branches in other cities in the early second
millennium BC. In fact, the very same archaeological excavations
have shown, and mapped with increasing detail, that production
remained decentralised and not just confined to a few centres or
large workshops. Ironically, Rostovtzeff himself, the leading pro-
ponent of the old-style modemism, saw precisely this as what
marked the limits to the modern nature of the Roman economy.
The technological and economic advantages to be gained from a
concentration of capital and production in large units were insuf-
ficient to produce and sustain an industrial development, modem
style. 43
Second, the nature of the market in the Roman world. Finley
tooktheseeminglyunproblematicalviewthattheancienteconomy
40 Whittaker 1995:22. 4I Kenrick 1993:36. SeealsoWilson2001.
42 Cf.Fulle 1997andPrachner1980ontheorganisationofltalicterrasigillataproduction,
stressing the limited scale of most potteries and the importance of travelling potters.
Furtheron the interpretationofGallicterrasigillataproduction, seethediscussionsby
Woolf 1998, chapter 7 and Whittaker2002. Seminal are Goudineau 1974 and Marsh
1981.
43 Rostovtzeff 1957: 172-179. See Larsen 1976 for the organisation of early Assyrian
merchants.
30
BEYOND THE ANCIENT ECONOMY?
was not organised as "an enormous conglomeration of interde-
pendent markets". Recently, however, the economist Peter Temin
argued that Finley was quite simply wrong, the economy of the
Roman Empire representedjust such a conglomeration.44 This is
an extraordinaryclaim. Onemight conceivablyimaginethat some
markets had begun to be linked by middle- and long-distance
trade. But to see the entire economy, spanning several continents,
as organised by a set of interlinked markets is quite another mat-
ter. It is doubtful whether the mature eighteenth-century European
economy, outside some restricted pockets, could be described in
such terms.45 Given our almost complete lack ofprice datawhich
could substantiate this conclusion, caution might be advisable. 46
But to Temin the mere occurrence of markets is enough to justify
the claim. "If about seventy-five per cent of the population of the
Roman Empire wasengaged in farming, then it is not unreasonable
to suppose that over half of production was carried on by house-
holding rather than by market exchanges", he observes undeterred.
For, according to Temin, the prevalence of households in organ-
ising economic activity does not mean that peasant production
was not determined by the market, only "that most of eachfarm s
activities were devoted to maintaining its workforce".47 This will
not do; it is an attempt to explain away the fact that for most
peasants, the market will only have constituted a subsidiary outlet
anddidnotprovide themwiththeirprimary meansof subsistence.
44 Temin200[: 181.
45 Cf. the cautious analyses ofPersson 1999 and Erdkamp 2005. See Osbome 1996 for
an interesting study of some link-up between market and production in the archaic age.
But that does not make the economy as a whole a conglomeration of interdependent
markets, cf. Morley 2007: 96.
46 A later MIT working paper by Kessler andTemin (2005) reasserts this claim of such
Empire-wide integration based on a good handful ofprice observations scattered across
several centuries. The scarcity ofobservations, however, only serves to reinforce the pat-
tem detected, the authors declare optimistically. That is methodologically unwarranted;
there is simply far too little data even to begin that form of analysis. An Empire-wide
study of market integration on the basis of price data is beyond our reach. At best,
and even that is debatable, the price data recovered from Egypt may allow us a broad
impression of conditions in one province. See Chapter 3 of this book for what can
hesitantly be achievedand with the opposite conclusion.
47 Temin 2001: 180. The observation that even in today's rich societies a large part of
economic activity is still kept within the family misses one crucial distinction. Most
middle-classhouseholdsdependonwageincomefortheirsubsistence,whereaspeasants
have historically produced their own subsistenceoutside the market.
31
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
A peasant household, therefore, cannot easily be abstracted to the
model of a profit-orientated firm, as Temin seems to be doing;
its primary objective was different, the reproduction of the family
unit. In other words, even on Temin's own premises more than half
of the Roman economy was not organised as "a conglomeration
of interdependent markets". The question of a market economy
in the Roman Empire is clearly less straightforward than Temin
allows.
On closer inspection the various modernising uses of concepts,
which have been explored here, seem to a large extent to have been
guided less by analytical than by value judgements. A clear indi-
cator is the recurrent use of terms such as "complex" and "sophis-
ticated", qualifying or supporting the descriptions. By themselves
these are almost empty words, meaning as little as "not primi-
tive" or "something not inferior". Struggling to free themselves
from the confident shadow of nineteenth-century evolutionism,
twentieth-century social anthropologists have taught us that there
is no such thing as a primitive culture, only imperceptive inves-
tigators. Hence words such as complex and sophisticated often
appear in texts mainly as apologetic or defensive value markers. A
remark made in connection with a discussion of ancient accounting
by historianWilliamHarris illustrates my point neatly. Againstthe
claim that the ancient agricultural writers did not show any signs
of modern economic rationality and capitalist accounting prac-
tices, the objection is voiced: "But Columella has in turn been
defended."49ObservationsthattheGreeksandRomansstructured
their thinking about estate management on other principles than
those at the core of modem accounting and management theory,
are taken as making an implicit statement about their cultural short-
comings and lack of rationality. But that is not the case; it only
means that they organised their thought in a fashion different from
some patterns which have been gaining ground since the birth of
modern capitalism.49 Essentially the problem is one of empathy.
To Weber and Finley, as to most anthropologists, the culturalist
4 Harris 1993: 25 (my emphasis). For a further example, see Rathbone 1991,
chapters 8-9.
49 Cf. Kehoe 1997; see Mickwitz 1937 for a classic analysis. See Tchernia 2005 for a
recent collection of papers dedicatedto this issue.
32
BEYOND THE ANCIENT ECONOMY'!
solution, Herder-style, was the most appealing. This meant under-
lining the differences between "us" and "them" and subsequently
seeking to understand the logic governing "the other" rather than
applying our own habitual notions of normalcy.
To the critics, however, it is exactly that approach which is
the central problem. Here are Peregrine Harden and Nicholas
Purcell: "The consequence has inevitably been to devalue practices
which cannot be usefully seen as embryonic modernity.. . For
the telescopic perspective encourages the itemizing of absences
and shortfalls ... Ultimately, however, the project was sabotaged
by the use of the 'minimalizing' comparative technique. Its trust
remained essentially negative: a generation was taught... what
was not important in the pre-medieval Mediterranean. "50 The con-
trasting comparisons between the ancient and the early modern
world, inherent in the Weberian approach, are (mis)perceived as a
derogatory portfolio of defects and flaws - in the Italian archae-
ologist Carandini's words "an annihilation of the past" and "self-
deification of the modem world".5' But it is difficult to see how
one could argue against a strongly modernist position, preferring
instead to emphasise the unique or historical characteristics of the
Greco-Roman world, without showing the contrasts and differ-
ences. That does not imply an essentially negative argument. On
the contrary, it is the central point in Weber's and Finley's analyses
that we find a world governed by other principles: the dominance
of agriculture, the dependence of urbanisation on aristocratic con-
sumption, the importance of slavery and the all-pervasive influ-
ence of an ideology of social status which privileged political over
entrepreneurial activities.
In any case, the current dissatisfaction with comparisons focus-
ing on the contrasts and differences between the Greco-Roman
world and later European developments has brought about a very
interesting change in the use of comparative history for the study of
the ancient economy. Pioneered by works such as D'Arms' Com-
merce and Social Standing, a number of scholars have taken up
the theoretical challenge of the "primitivist" position and begun in
a more systematic fashion to re-examine the relationship between
Harden and Purcell 2000: 146. 5I Carandini l983a: 178-179.
33
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
theeconomy oftheRomanEmpireandthatofmedievalandearly
modem, capitalist Europe. Rather than differences and contrasts, it
is now similarities and analogies which are sought out and empha-
sised.52 Observing that the Empire contained a number of very
large cities, with the one million inhabitants of Rome at the apex,
they justly note that only at a very late stage was the European
economy able to equal that achievement. London, the largest city
intheeighteenthcentury,didnotreacha millioninhabitantsbefore
l8oo. Hence, for instance, they warn againstpostulating any sig-
nificant difference in the size of the market. Instead, they take their
lead from the many developments in the study ofthe pre-industrial,
still agrarian-based,Europeaneconomy,emphasisingtheslowness
and limited extent of change. In the larger perspective, this was
a world, in the catchphrases coined by the French Annales histo-
rians, of la longue duree or histoire immobile, where economic
change tended to be cyclical, caught as it was in the Malthusian
trap where productivity gains in agriculture were underminedby
growingpopulation.Thecrucialdevelopmentfirstcameaboutwith
the Industrial Revolution around 1800. Until then the productive
and technological base of society was organic and set clear limits
to growth. Faced with this more "primitive" or static version of
early modern Europe, the Dutch historian and epigraphistHarry
Picketconcludesthatit isnowtime "finallyto call into questionthe
view of the primitivists that the world of the Roman Empire was
so very different from later pre-industrial Europe as they claim. In
this ... I would like to suggest that the economy of the Roman
Empire was more closely related to the economy of pre-industrial
Europeuntil the eighteenthcentury thanour owntime. "53
52 The approach was pioneered in studies of e. g. D'Arms 1981 and Schleich 1983. By
far the best and most important specimen is Picket 1990 which is a systematic attempt
to produce a synthesis of the Roman economy by analogy to early modern Europe
(1400-1800).Otherexamples includeHarris 1993; van Nijf 1997;Laurence 1998 and
1999; Harris 2000; Harden and Purcell 2000; and Temin 2001.
53 "endlich die Auffassung der 'Primitivisten' in frage stellen, dass sich die Welt des
Romischen Reiches so wesentlich van der spateren vorindustriellen Europa under-
scheidewiebehauptet.Indem. ..mochteichdieVermutungaussern,dassdieWirtschaft
des RomischenReichesmit der Okonomie des vorindustriellen Europabis zum 18. Jh.
vielleicht enger verbunden war als mit unserer Gegenwart" (my translation). Picket
1990:28.
34
BEYOND THE ANCIENT ECONOMY?
It is difficult to disagree strongly with Picket's position, stated
in such very general terms. It is certainly salutary to be reminded
in our comparisons that the scale anddimensionsof pre-industrial
economies were of another, much smaller order than in the mod-
ern world, that they, so to speak, belong on the same baseline. 54
However, this is not where the main argument is. Even Finley, for
instance, emphasisedthat urbanisationin Greco-Roman antiquity
reached a historical high point, only equalled very late in history.
He also expressed strong doubts about relying too much on analo-
gies furnished by anthropological studies of, say, gift-giving in
much simpler societies. The market as a form of exchange was
far too important in antiquity; and no one who has walked around
the Monte Testaccio - the man-made, almost eighth hill of Rome
consisting (for the most part) of oil-bearing amphorae discarded
in the days of the Empire - can seriously doubt the existence of
quite large volumes of exchangewithin the RomanEmpire.55 But
Picket takes the analysis much further, constructing the Roman
Empire closely on the model of early modern Europe. Thus a pic-
ture emerges where regions of relative low-level agriculture and
only sparse urbanisation intermingle with dynamic, more modem
centres dominated by highly specialised agriculture in the Dutch
style, and towns relying more on trade and export manufacture for
their survival than on aristocratic consumption. Other contribu-
tions to the debate point in the same direction when they suggest
thatthe Roman Empireshouldbeseenas a marketeconomy work-
ing essentially at a level comparable to that of the Atlantic world
in the eighteenth century. 56
The problem with such analyses is that they depend on a very
loose use of the notion of functional equivalents. Being strongly
disinclinedevento considerthepossibleimportanceofdifferences,
they tendto prefereitherto declarethemmarginalor automatically
54 Nicolet 1988: 37-38 and e. g. Andreau i994b: 191 and 1995: 948 point in the same
direction.
55 Finley 1970, 1975,chapter6 and1977,forinstance.OnMonteTestaccio,seeRodnguez-
Almeida 1984; Blazquez Martmez and Remesal Rodriguez 1999-2001.
5 Temin 2001. Also Harris 1993:19.
35
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
assert that "the methods may have been different, but the economic
consequences are the same". 57 A discussion of the rank-size dis-
tribution of Roman towns is symptomatic. Referring to De Vries'
influential study ofurbanisation and urban networks in early mod-
ern Europe, Picket concludes that the rank-size distributions of
Roman cities can be seen as roughly similar, even if admittedly
there are some differences, especially among the very large cities
of the Roman and early modem world. In De Vries' study, how-
ever, it is a central point that the shapeof urban systems is closely
linked to the specificsocio-economicstructure ofanygivenpartic-
ular society. Differencesat the top of urban hierarchies,therefore,
oughtnotto be quickly brushedasideasofonly trivial importance;
they may reflect significant differences in economic organisation
between the two societies. 5 Braudel, after all, argued that capital-
ism was a higher-orderphenomenon,connectedprecisely with the
top of societal hierarchies in early modern Europe rather than the
virtually immovable mass of peasant producers and small towns. 59
In Picket'sanalysis,however, this type of issueis simply bypassed
in favour of a broader, more inclusive understanding of "mod-
ern" economic phenomena. But there is a price to be paid for
such a manoeuvre: a potential loss of explanatory power. Ulti-
mately, the economies of early modern Europe and the Roman
Empire followed very different paths of development. The more
we downplay the differences, the less able are we to understand
why.Thatis unsatisfactory.The aim shouldbeto findbetterhistor-
ical models to help us elucidatetheparticular natureofthe Roman
economy.61
57 Picket 1990: 39: "die Methoden mogen verschieden sein, aber die Wirtschaftlichen
Folgen sinddieselben"(my Englishtranslation). The method waspioneeredby D'Arms
1981:165-169.
58 See Woolf 1997 for a discussion linking these differences to the presence of a single
unifying empire in the Romanworld.
59 Braudel 1982:455-457.
60 Picket 1990: 146. Similar in style is Harris 1993: 24, advising Love 1991 to adopt a
more inclusive definitionof capitalism so as to make room for antiquity. De Vries 1990
explains his approachto urban networks.
' For a similar wish to transcend this problem cf. the introductions to Mattingly and
Salmon 2001 and Manning and Morris 2005.
36
BEYOND THE ANCIENT ECONOMY'7
From traditional Asia to early modern Eurasia
At first sight the persistent attraction of early modern Europe
as yardstick for the ancient world appears curious. The last half
century has seen a strong challenge building up to the notion of
European developments as the quintessential expression of history.
Europe's paradigmatic status in world history has suffered consid-
erable erosion and is no longer self-evident. Nonetheless, the
current modernising trends in ancient history are not an isolated
phenomenon. Similar doubts about the legitimacy of analysing
differences between "the West and the rest" have occurred in a
whole range of disciplines. During the 1990S a rather curious
debate took place between Marshall Sahlins and Gananath Obey-
sekere, both distinguished American anthropologists, about the
proper interpretation of Captain James Cook's death on Hawaii on
his second return, 14 Febmary 1779. To Sahlins the explanation
has to be found in the particular cognitive and cultural universe
of the Hawaiians. Unfortunately for Cook, his arrival was linked
to their calendar of religious ritual by the Hawaiians, who saw
him as their returning god Lono. Cook willingly played along, not
knowing that the festival part of the year would end in the king
re-establishing his primacy by appropriating the powers of Lono.
Even then nothing would have happened if Cook had not been
forced by a broken mast to return after first having left the island.
This action was perceived as a threat to the cosmic order where
the king was now in command. That order had to be re-created
and Cook/Lono was killed so that his powers could properly be
passedto the king. To Obeysekeresucha culturalist interpretation
is the result of blatant racism, making a mockery of the cogni-
tive abilities of the indigenous population and denying them their
rightful place within common humanity; of course, the Hawaiians
could tell the difference between a white explorer and their god.
Instead he wants to render fuzzy the clear distinctions between the
visiting Westerner and the indigenous culture. One cannot avoid
62 Hodgson in 1974 and 1993, presents one of the most thoughtful attempts to re-evaluate
the position of Europe in world history.
37
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
West. Indeed, how could it possibly be in the present world? The
legacy of European imperialism is a heavy one and the romanticist
emphasis on the native cultural tradition has more often than not,
quite understandably, been combined with a wish to gain access
to the technological and economic resources of modernity. The
scenes which one could see some years ago of Hindu nationalists
celebrating in the streets of Delhi because India had now joined the
atomic powers are only an extreme illustration of a more general
condition.66 In his essays on Istanbul, Orhan Pamuk, the great
novelist, has managed to capture well the tortured and complex
relationship between Westernising and "othering" in the cultural
identity of third world countries:
The viciouscycle is fed by Westemisingintellectuals wholongto hearthepromi-
nent writers and publishers of the West praise them for being like Westerners.
Writers like Pierre Loti, by contrast, make no secret of loving Istanbul and the
Turkish people for the opposite reason: for the preservation of their Eastern par-
ticularity and their resistance to things Western. In the days when Pierre Loti was
criticising Istanbullus for losing touch with their traditions, he had only a small
following in Turkey, most of this ironically, among the Westemising minority.
But whenever the nation is embroiled in an international dispute, the Westemised
literary elite makes an indignant peace with the "Turkophilism" of Pierre Loti's
highly sentimental andexoticwritings.67
The position of being "other" both holds out a promise of autonomy
and a danger of falling into the trap of the merely quaint and exotic,
an object of domination.
This problem of "empowerment" was accentuated by the pro-
cess of de-colonisation in the decades following the Second World
War. No longer was it possible for historians to regard the Western
world as the sole agent of social development, the faber histo-
riae, when it now became clear that history did not end with the
European colonial empires. Hence there has been a growing dissat-
isfaction with the traditional view of world history as the unfold-
ing drama of Western development from the birth of civilisation in
Mesopotamia through antiquity, the Middle Ages and Renaissance
to modernity, leaving the rest of the world in the sidings until it
EvenHuntington , chapters 4-5 sees how "indigenisation" and modernisation may
go together.
Pamuk 2005: 213.
40
BEYOND THE ANCIENT ECONOMY?
came under the influence of Europe. Instead the main concern has
been to establish a more global point of view and, in the words of
Eric Wolf's influential book, include The People Without History.
One of the most remarkable results of these endeavours is the
American Islamist Marshall Hodgson's great work of comparative
history The Venture of Islam. Hodgson asked how world history
should be conceived before the establishment of European world
hegemony. His answer was to relegate Western feudalism to the
margins of the historical narrative. The dominant force in the world
from the seventh to the seventeenth century was, he observed, the
civilisation of Islam. Europe need not, in other words, be taken as
paradigmaticofhistory as such.68
In the study of India and the Indian Ocean economy Hodgson's
analysis was foreshadowedby the Dutch historianJ. C. van Leur.
In the 1930s he began to reject the notion of European superiority
right from the arrival of Vasco da Gama in the early sixteenth
century. European hegemony was a much later development, van
Leur maintained; during the first centuries of their presence the
influence and importance ofthe Europeans were only marginal and
certainly amounted to far less than commercial, let alone political
control. This is how he drew the larger picture:
Such an East, retiring into isolation, standing to one side along the path of the
West, the path ofhurried activity, world trade, world economy, modem life, seems
to have been foreign to the eighteenth century. That century did not know any
superior Occident,nor any self-isolatingOrientno longerprogressingwith it. It
knew a mighty East, a rich fabric of a strong, broad weave with a more fragile
Western warp thread inserted in it at broad intervals. 69
For a long period after the Portuguese first establishedtheir sea
empire in the Indian Ocean, Asia remained largely unaffected and
followed its own course of history. The application of European
periodisations was therefore meaningless.
At a general level, van Leur's ideas have formed the basis of a
newparadigmwhichhasbeendeveloping sincethe SecondWorld
War. One important implication of his view, which has emerged
withevergreaterclaritysincethe 1950s,istheexistenceofa strong,
Hodgson 1974.SeefurtherHodgson1993:44-71 and91-94.
9 Van Leur 1955: 289.
41

THE ROMAN EMPIRE AND COMPARATIVE STUDY
independent. South and South-East Asian trading world and mer-
chant community before the beginning of European colonialism
from themiddleoftheeighteenthcentury. 70The Asianmerchants
proved to behighly competitive. Only in a certain restricted num-
ber of trades where the European chartered companies were able
to impose their own terms ofcommerce didthey lose out. But for a
long period they held their own in most other fields, restricting the
influence of the Europeans to that of yet another participant in the
already existing trading system. 71 In that respect, van Leur's ana-
lysis wasonly thebeginning. Stressing, like Finley,theclearlimits
of commercial development in the pre-modern world, he empha-
sisedtheephemeral position oftradewithin"thehighlytraditional
agrarianworldstill basedcompletely onself-sufficiency, withlittle
independent city life" and "closed village communities as social
and economic basis". 72 However, as the resilience of Indian and
other Asian merchants has been documented, the emphasis has
changed markedly and is now placed rather on the large amounts
of trade and the vigorous commercial developments in that part
of the world. As the controversial historian of early modern trade,
Niels Steensgaard, hasremarked, it isasifthemyth ofthe immense
riches ofAsiahasbeen substitutedfor that ofthe static, traditional
Orient.73 Even so, the foundation of van Leur's judgement has
suffered some erosion. The study of the allegedly autarkicIndian
village communities, the archetype of traditional society, has all
but gone through a revolution. It is now clear that these villages
were not entirely closed economies but also entered into wider
marketing networks. For the Mughal Empire, its leading historian
Irfan Habib has identified a long period of commercialisation of
the agricultural surplus provoked by the tax demands of the state.
Thishasbeenfollowedby seminalworksofscholarssuchasB.R.
Grover, K. N. Chaudhuriand C. A. Bayly, linking the rural world
7° Braudel 1984;Arasaratnam1990; Marshall 1993;Wills 1993;andDisney1995:484-
535 areconvenientandshortoverviews.Formoredetailedtreatments, seeChaudhuri
1985; 1990 and Das Gupta and Pearson 1987.
71 Steensgaard 19733: 405-41 (, though often misunderstood, is the best analysis of the
resilienceofAsianmerchants,togetherwithDasGupta 1979;Prakash1985;andBaner-
jee 1990.
72 Van Leur 1955, quotations from 283 and 282. 73 Steensgaarcl 1987.
42
BEYOND THE ANCIENT ECONOMY'?
of northern India in the seventeenth and eighteenth centuries to a
hierarchy of markets. 74
Similar developments have taken place in other branches of
Asian or Oriental studies. In 1966 Maxime Rodinson coun-
tered the widespread assumptions about the cultural blockage of
Islam towards capitalism. In his analysis, he objected that even
Mohammed had been a trader and that the religion was, in many
ways, quite favourable towards mercantile activities; Islam and
capitalism were not mutually incompatible. The following year,
Goitein publishedthe first volume of his extraordinary account of
a community of Jewish traders centred in medieval Cairo, show-
ing a strong andhighly developed commercial world. In Ottoman
studies, the picture of a stagnant and inefficient empire in the
plunderinghandsofan Oriental despothasbeendiscarded.Before
it became "the sick man of Europe" the Ottoman Empire had
been a strong and equal adversary with a flourishing economy,
a highly developed market sector and capital formation.75 Then
there is China, where research seems to defy everything we take
for grantedaboutpre-industrialsocieties.JosephNeedhamandhis
team of assistants, for instance, have identified a strong scientific
tradition under the various Chinese imperial dynasties, in many
ways the equal of Greek and, for a long time, European science.
This is to say nothing of the spectacular rise in steel production
underthemedievalSongdynasty,not significantlysurpassedagain
before the Industrial Revolution, or the large internal rural com-
mercial developments under Ming and Ch'ing, all brought to our
attention by scholars such as Elvin, Hartwell and Skinner.76
The result of all this has been a strong challenge to our very
notion of modernity. As it has become increasingly clear that
the concept of traditional society is difficult actually to locate
empirically in the pre-industrial world, even in the archetypal case
of India, the very notion of differences between early modem
Europe and the other complex civilisations of Eurasia has come
74 Habib (963; Graver 1966;Chaudhuri 1978;Bayly 1983.See Inden 1990:chapter4 on
Indian village economies.
75 Inalcik 1969; Faroqhi 1991; Inalcik and Quataert 1994; Atfanir 1997. Bj0rkelo 1999 is
a convenient survey.
7 Some of their more optimistic conclusions may be in need of moderation. For some
cautionary critiques, see Finlay 2000 and Wagner2001.
43
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
under strong attack. In the now classic, if controversial, Oriental-
ism (1978) and the sequel Culture andImperialism(1993) it was
argued by Edward Said that the differences were but a romantic
fiction of the Europeanmind, longing for the exotic. This was in
the best case. In the worst, they were an imperialist construction
designed to dominate the colonised societies. Thus static tradi-
tional society with all its connotations ofcommunal heart-warming
Gemeinschaft and lack of commercial activity had not been there
before. Rather, it was imposed on the dynamicAsian societiesby
the colonialist regimes - a case of the development of underde-
velopment. 77 Understandably, we are now called upon to undo the
wrong and re-admit these societies into modernity. Here is Sanjay
Subrahmanyam, a prominent historian of South Asian economies:
"Having taken away so much from the societies of South Asia, it
seems to be high time that social science at least gave them back
what they had by the sixteenth and seventeenth centuries - their
admittedly very ambiguous 'early modernity'. "78
If the developments in the study of the economy of the Roman
Empire are added, an image gradually emerges of a very old pre-
industrial Eurasian world of comparable high cultures. Lately,
there have been a number of attempts to synthesise this condi-
tion. Most radically, some have posited the existence of an all-
embracing world system of long-distance trade across Eurasia
since the Bronze Age, determining the fate of its constituent soci-
eties. However,thatisanextremeposition,consideringthelimited
amounts of trade passing across the Eurasian land mass. 79 Rather
more challenging are the recent attempts to reinvigorate a diffu-
sionist approach to the study of historical societies. So far, the
CambridgeanthropologistJackGoodyhasprovidedthemost sys-
tematically worked-through version of this line of thought. To
him the great divide does not run between Europe and Asia but
between Eurasia and black Africa. The latter never shared in the
Bronze Age urban revolution and this explains the problems of the
77 Wolf1982isthemostsystematicattempttodevelopthisidea.Morenuanceddiscussions
of Indiacanbe found in Bayly 1988:chapter5; Breckenridgeand van der Veer 1993;
and Bayly 1999.
78 Subrahmanyam1998: loo and l990b. Seealso Pomeranz2000.
79 Frank and Gills [993 and Frank 1998, inspired by Schneider 1977 and Abu-Lughod
1989.SeeSubrahmanyam 1996:xv-xvi for a Scepticalnote.
44
BEYOND THE ANCIENT ECONOMY?
continent in modernising today. The many, presently, much more
successful Asian economies can, on the other hand. be seen as
the result of a much greater infrastructural compatibility between
EastandWestwhichultimately datesbackto common roots in the
Bronze Age and the rise of civilisation. This shared legacy consists
in the development of similar socio-economic institutions which
Goody summarises as social stratification, literacy, specialisation,
division of labour and a mercantile capitalism.80 The result was
a series of complex and broadly similar pre-industrial agrarian
economies shaped by Smithian market dynamics and Malthusian
constraints, to borrow a set of concepts developed within late
imperial Chinese history by R. Bin Wong and more extensively by
Kenneth Pomeranz. 81
From "primitive" capitalism to agrarian empire
FIowever, it is difficult not to feel a little uneasy about this attempt
to inscribe capitalism into the DNA string of civilisation as such.
There is a very real danger of doing away with most of human
history, lumping it together under the same heading as a mir-
ror image of early modern Europe: a kind of Eurocentrism by
the backdoor. Of course, on a very broad, minimal, definition
such as the existence of market trade, generating Marx's famous
sequence ofM(oney)-C(ommodity)-M(oney), capitalism probably
has to be seen as ubiquitous in civilised societies. However, market
trade already provides conceptual coverage for that phenomenon.
Capitalism, on the other hand, will always remain linked to the
developments culminating in the Industrial Revolution. True, the
link was a complex one rather than a linear progression, as E. A.
Wrigley insisted two decades ago. The Netherlands, which had
spearheaded economic developments in the seventeenth century,
fell behind during the eighteenth century and failed to make the
transition to coal and steam till relatively late. The Dutch, though,
80 Goody 1990; 1996; 1998: l-26. Perlin 1983; 1993.
81 Pomeranz 200odeveloping anideaofWong 1997:Part (. FurtherGoldstone 1998:2002.
SeeParthasarathi2002 for a critique ofPomeranz,pointingout that he underestimates
the importanceofEuropeanmarketdevelopments andstatepolicies.
45
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
had begun to use peat on an extensive scale for production pur-
poses, but their industries lost out to protectionist competition
from French and British manufacturers.82 Capitalism was gradu-
ally coming together to drive economic development in a system
which transcended individual countries and produced a dynamic
succession of leading economies, as Braudel described it in the
final volume ofhis great work on capitalism. Why these develop-
ments took place in early modern Europe rather than somewhere
else is a question that cannot be avoided, indeed should not, con-
sidering their world-shattering effects. Of course, there is a risk
of seeming Eurocentric in providing answers. But that should be
treated as a challenge, not as an excuse to brush aside the issue.
Thetaskconfronting us isto provide explanations ofearly modem
European developments without denying the existence of histori-
caldynamicsinotherculturesandtimes. 83HereneitherGoodyand
Pomeranz. nor Picket, with their strong determination to identify
analogies and parallel patterns, have much to say. Yet, if the basic
tenor of their argument is correct, suggesting answers seems more
necessary than ever, unless we have given up trying to understand
the different trajectories of world history. To cite a recent col-
lection of papers, we need a kind of historiography "committed
neither to justifying European domination, nor to explaining it
away".84
82
Wrigley 1988. Jongman 2006 takes Wrigley to show deye lopments
that the economic
preceding industrialisation wereunimportant, andthatthereistherefore little todistin-
guishearlymoderncapitalism.WhileI agreewithJongmanthatweneednotshapeour
analyses of the ancient economy on the early modem model, his reading of Wrigley
is an overinterpretation. Wrigleydescribestheperiodof vigorouspre-industrialeco-
nomicdevelopment as a necessaryprecondition,but not an automaticguaranteefor
an industrial breakthrough (l988:'ll-l2, 113-118). Even so, Wrigley's vision of the
processremainsheavilyAnglocentricandpayslittleattentiontotheoveralldynamicof
theemerging capitalist system. Thatwould havecome out moreclearly ifhehadalso
includedprocessesofstateformationandmilitarycompetition,cf.McNeill1982(mil-
itarycompetition);Wong 1997:chapter6 (comparativedifferencesin stateformation
betweenEuropeandChina);DeVriesandWoude1997(onthemodemcharacterofthe
Dutch economy).
83 OnthislatteraspectI agreewithMorley2007:6-9,butthatdoesnotrequirethestudent
ofancienttradetoplaydownthedifferencestoearlymoderncapitalism Landes1998:
415-418 is a pointed, 'if perhaps too strong, critique ofthis basically Saidian tendency
to ignoreortonedownthequestionofdifferences Pearson 1991:45;Wong 1997: 1-8,
andBayly2000insistonthecontinuedneedtodifferentiateourimageofpre-industrial
Eurasia.
84 Manning and Morris 2005: 9.
46
BEYOND THE ANCIENT ECONOMY?
I donot advocateintroducinga banoncomparisonsbetweenthe
Roman Empire and early modern Europe. Recently, for instance,
attention has been called to a possible similarity in the effects
that eighteenth-century London and early imperial Rome had on
their respective hinterlands. The two capital cities were of com-
parable size and both represented an enormous concentration of
spending power; the effects of Rome on her hinterland might have
been parallel to those of London on hers, where the rising demand
of the city generated economic development in the countryside,
marked by rising productivity and an increased commercialisation
of agriculture. 85 But we also need to explore the limits to this
comparison. The enormous spendingpower of the Roman market
was the result of the concentration in that city of a large part of the
resources generated by a territorial empire spanning the Mediter-
raneanworld with a population ofperhaps 60 million. In 1750, on
the other hand, when London's population reached 700,000, she
still had to rely mainly on the British Isles and a few commercial
colonies and outposts around the world, in all comprising a popu-
lation in the order of magnitude of say 10 million. In comparison
with the ancient imperial metropolis, London's resource base was
much smaller. The English capital had to compensate by making
much larger earningsfrom foreign trade than did Rome to finance
its consumption.86
The relationship between empire, concentration of spending
power and commercialisation-cum-growth was first brought to the
forefront of the debate on the economy of the Roman Empire by
Keith Hopkins, three decades ago. 87Taking his point ofdeparture,
like Weber and Finley, from the politically defined consumption
patterns of the agricultural surplus, he drew attention to the way
these were modified by the creation of large empires. 88 A part of
the resources which had formerly been available to the multitude
of local cities around the Greco-Roman world was now confis-
catedandredirectedto serve the purposes of the Roman imperial
85 Hopkins 1995-6:59;Morley 1996.
86 PaceMorley 1996:29,whotendstounderestimatethesignificanceofthisdifference.
SeeWrigley [987: chapter6 andReed 1996for theeconomyofLondon.
87 Hopkins l978a;1980;19830;l983b;and 1995-6.
SeenowalsoErdkamp2001onthesimilaritiesofthe"consumercity"andHopkins'
model.
47

THE ROMAN EMPIRE AND COMPARATIVE STUDY
authorities. Basically, that meant that they were channelled to
Rome, the capital of the Empire, and the army, stationed along
the frontiers. Since many of these taxes were, presumably, paid in
coin or at some stage after delivery in kind, sold on the market and
converted into money, most areas of the Empire where there was
no government spending would have hadto earn their tax money
backif they were goingto continue to beableto paytheirannual
taxes and not run out of coin. The result was a limited growth and
commercialisation process where taxpayers around the Empire
entered the market to earn their tax money, and export flows of
manufactured products arose between the areas in tax deficit and
those in surplus.
Hopkins' model hasjustly attracted much attention. Not surpris-
ingly, 'giventhecurrent climate ofdiscussion, many haveempha-
sised the elements dealing with growth and commercialisation,
whereas the limits to these processes and the close affinity of the
model with that of "the consumercity" tend to be ignoredor dis-
missed. 89Hopkins hasevenbeencriticised forunderestimating the
amount oftrade and the penetration of market forces in the Roman
Mediterranean. 90 Here, Harry Picket's endorsement and the spe-
cial twist he gives to the taxes-and-trade model are of particular
interest. Rejecting itsWeberianfoundations, hesetsoutto compare
thetax system oftheRomanEmpirewiththatoftheearlymodem
European nation states. He stops atthe need ofthe latter to borrow
from private financiers and the gradual creation of a permanent,
funded government debt. This institution is wholly lacking in the
Roman Empire. In Picket's view, this is a reflection ofthe greater
strength ofthe imperial state; it did not need to resort to such
measuresto finance'itsactivities. It could simply rely on taxingits
subjectsdirectly. Fromaneconomicperspective, however,thetwo
fiscal regimes would have produced similar results. In both cases
the state absorbedthe wealth of its citizens and created a similar
amount or kind of economic activity through its consumption. 91
8t) SeeErdkamp2001 for a similarview of the"consumercity .
90 SeePaterson1998:153andPicket1990forcritiquealongtheselines.Greene1986isa
typicalexampleemphasisingcommercialisationandgrowth.Themodelhasalsobeen
criticised from more "primitivist" perspectives, e. g. Jongman 1988; Whittaker 1989;
Duncan-Jones 1990; and Whittaker 1995.
'" Picket 1990: 39 and 46-47.
48
BEYOND THE ANCIENT ECONOMY?
I do not share the confidence of this conjecture in its reliance
on functional equivalents. A basic, if rarely noted, premise of
Hopkins' argument is that there is no such thing as the market,
in the abstract general sense. It is not just out there, happening
on its own. Rather, it needs to be created and often even pressed
or forced into existence under pre-industrial conditions. Hence it
becomes crucial to understand how the particular markets are artic-
ulated, howthey are moulded and shapedby the social andpolitical
institutions of society and the context in which they appear. This,
of course, was Polanyi's point and the one that has survived the
many empirical criticisms of his economic models. It has even
entered the mainstream of economics through the institutional
analysis of DouglassNorth.92 In this particular instance, the rela-
tive weaknessoftherising nation state andits consequentreliance
onprivateloanscannotjust bereducedto a matterofconsumption
in relation to the economy. A main result of that situation was
the formation of an efficient international market in capital. The
development began when the large medieval fairs of Europe were
turned into meeting places for the rich banking firms, connected
with the European monarchs, where they could settle accounts and
deal in bills of exchange. Gradually, as the scale of transactions
increased, more permanent structures became necessary and in
the seventeenth century this function was taken over by the stock
exchanges of Amsterdam and London. During the same period
the government debts of the Netherlands and England were put
on a more stable basis in the form of negotiable, interest-bearing,
bonds. Thus this line of investment was now effectively opened up
to groups other than the closed cliques of leading financiers. The
result was an important extension of investment opportunities for
commercial capital. 93 The point is that such an institutional devel-
opment seems to be largely absent from the Roman Empire. We
do find a hierarchy of periodic fairs and markets. However, they
remained, to a very large extent, gathering places where the rural
population could sell its surplus production and have its limited
92 Steensgaard 1981; 1984 and Lane 1966: 383-428 are some of the most perceptive
analyses of the close bonds between markets and the socio-political formation. See
further Polanyi, Arensberg and Pearson 1957; North 1990; and Morris 2002.
93 De Vries 1976: chapter7 and Braudel 1982: 8i-l 14.
49
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
needs for market services supplied. The larger fairs also seem to
have played a role in regional and interregional trade. A few even
developed into important foci of long-distance trade, such as the
one on Delos. But they do not appear to have fulfilled the role of
"international"credit markets. We should, perhaps,ratherturn our
attention to the cities. Here we do, of course, find market build-
ings and squares, but there is nothing to suggest that they would
have operated on lines resembling those of the formalised stock
exchanges of seventeenth-century Amsterdam and London. What
is more, on Picket's interpretation of the taxes-and-trade model,
stressing the absence of government debt, we would not expect
this either.94
Nor should we, asPicket concedes, expect any kind ofbourgeois
developmentgivingriseto a classoflargemerchantsandfinanciers
able to challenge the economic preponderance of the political elite
based in landed wealth. However, that cannot just be written off as
a difference in social structure. It had economic consequences as
well. It would, for instance, be difficult to imagine a situation like
that described by Braudel for the West Indies, where the plantation
ownersoften hadlargeproblems makingendsmeet whilethe mer-
chants who bought their produce and sold it back in Europe made
huge profits. 95 Indeed, in the Roman Empire things took place in
the opposite fashion. Wealthy landowners remained firmly in the
driving seat. In parting with their produce, they seem generally
to have been able to shift most of the risk onto the commercial
middlemen, to judge from the business procedures envisaged in
legal regulations and prevailing forms of contract. 96 This impres-
sion is confirmed by what we know about elite mobility in the
Empire. The crucial component seems to have been access to large
agricultural resources; it was the cities with the largest and rich-
est countrysides available that proved most successful in getting
members of provincial elites promoted into the imperial aristoc-
racy. The spectacular rise of members of the elite of North African
Lepcis Magna, culminating in the Severan emperors, should not
94 Finley 19853: 195-196,confirmedby deLigt 1993andFrayn 1993.OnDelos,seeRauh
1993- See Andreau 1999: chapter 11, pace Harris 2006, for the relative weakness of the
financial "sector" in the Roman economy.
95 Braudel 1982: 272-280. 96 Morley 1996: [59-166.
50
BEYOND THE ANCIENT ECONOMY?
be seen as the result of profits generated from the administration
of Mediterranean trade. Rather, their success came about because
they controlled an unusually large hinterland which provided them
with a vast agricultural surplus, especially in olive oil, that could
be exported.97 In other words, Picket tends to ignore the fact that
differences in social organisation produce the articulation or dom-
inance of very different economic interests. The Roman Empire
may, on very rare occasions, have allowed considerations about
the safety of its traders against encroachments in foreign countries
to enter into decisions about war and peace. But that is a very
far cry from the active strategies of Dutch merchants in the seven-
teenth century, successfully waging war around the globe to secure
and conquer markets. Here commerce was the dominant motive.
Could we imagine the Roman Empire in a situation like that of the
Spanish Habsburgs fighting on and off against the Dutch rebels for
eighty years and in the end depending on the very same Dutchmen
to provide credit to finance their armies? It all came to an end in
1648 when the Dutch merchants finally decided that peace was
more profitable than war, and shut their coffers. 98
In a historical perspective this early modern European develop-
ment, often described in the convenient shorthand ofmercantilism,
appears to be very much the exception, a unique occurrence."
Adam Smith sensed this clearly when he chose in The Wealth of
Nations to portray the preceding centuries of European economic
history as a perversion of the natural hierarchy between agricul-
ture, manufacture and commerce. The so-called mercantile sys-
tem had allowed the merchants too much power and far too many
privileges, and had made it possible for them to appropriate and
control a disproportionately large part of the total wealth of soci-
ety to the detriment of the overall well-being of the economy. This
should, according to Adam Smith, now be brought to an end; the
leading position of agricultural production should be recognised
and mercantilist policies abandoned. The principle of laissez-faire
97 See Mattingly I988b on LepcisMagna's aristocracy; Syme 1977 and Corbier 1991:223
for the situation in general.
98 Burke 1988: 229-230. On Dutch trade, see Israel 1989. See to Cascio 1991 and
Andreau l994a and 2003 for the relationship between market and government in the
Roman Empire. Harris 2003 effectively confirms their description.
99 Gellner l988b; Pearson 1991.
51
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
would then operate to restore the proper and economically most
beneficial relationship between the different sectors of the econ-
omy. '00 We are a far cry from the horizons of modern-day free
trade enthusiasts. Indeed, it is striking when one reads through the
pages of The Wealth of Nations how Smith's vision of the Chinese
Empire, resting on the backs of a multitude of hard-working peas-
ants and a large internal market, in some respects came quite close
to his ideal of the natural relationship between agriculture and
trade. 101
China may even be much closer to the historical norm than the
liberal Smith, who also saw oppression and consequent distortion
of the economy caused by the political elite, would have liked to
admit. In post-war scholarship the Chinese Empire has, in analyti-
cal contrast to Europe, come to represent the more normal pattern
of development, where a political elite, depending on agricultural
rent-taking, is able to dominate the commercial groups of society.
The locus classicus ofthis literature is a counterfactual explanation
of why China, clearly the leading economic power in the world
around 1400, eventually lost its lead to Europe. 102 In the thirteenth
century a brisk trade had developed in South East Asia, dominated
by Chinese merchants. But when the Ming dynasty took over from
the Mongols in 1368 this foreign trade was prohibited. Instead
the authorities tried to control contacts with the world outside and
attempted to organise foreign exchange relations within a large
imperial tribute system. As a consequence a number of very large
naval expeditions, headed by the legendary eunuch admiral Zheng
He, were sent out from China between 1405 and 1433, navigating
the South China Sea and the Indian Ocean, even reaching the coast
'"° Smith 1976: 276-85, 360-427 and Book 4. Smith's analysis lives on in the new eco-
nomic history which downplays the significance of foreign trade to eighteenth-century
Britain, e. g. Wrigley [987 and 1988. The argument ignores the contribution of a much
strengthened commercial infrastructure (institutions), that is simply taken for granted,
cf. O'Brien 1998.
101 It isno coincidencethatElvin'sclassicanalysisoflate imperial China(1973),describing
thecountry ascaughtin a high-levelequilibriumtrap, is rightout ofthepagesofSmith
(as is Pomeranz 2000). Smith expected something similar to happen for England,
Wrigley 1988.
'"2 McNeill 1982: chapter 2. See further Wallerstein l974a; Jones 1981; 1988;and Landes
1998.
52
BEYOND THE ANCIENT ECONOMY?
of East Africa. 103 The scale of operations was truly magnificent.
The first expedition is believed to have been made up of some
28,000 men and perhaps up to 255 ships of varying sizes, inclucl-
ing 62 enormous treasure ships. This was far beyond anything the
small European states at that time could even dream of achieving.
The handful of Portuguese ships that began to arrive in the Indian
Ocean at the end of the fifteenth century would have been no match
for these gigantic fleets. This has led a number of scholars to ask
why it was not the Chinese who sailed around Africa and began to
establish colonies in Europe.
As it happened, the Chinese had already aborted their expe-
ditions when the Portuguese arrived. A mounting Mongol threat
on the inner Asian frontier, a war in Vietnam and the building of
Beijing as the new capital strained imperial finances. In 1421 the
undertakingof newexpeditionswascalled offtemporarily. A final
voyage was completed between 1431 and 1433 before the project
was finally abandoned. In 1436 the building of new sea-going ves-
sels was prohibited, and the crews and resources were channelled
into traffic on theGrandCanal andpublic buildingactivities. After
a few years the know-how necessary to resume the expeditions
had been lost. The scene was set for the Portuguese and other
Europeans to embark on the slow and very gradual road towards
dominance. This is as far as the counterfactual will take us. Nor-
mally, the ills of despotism or Confucian dislike of eunuchs and
trade are called upon to explainthis historical "mistake". A closer
look, however, reveals that the very size of the Chinese expe-
ditions marks a very significant difference from the much more
commercially motivated European fleets. There was no way that
the stupendously costly activities of Zheng He could ever have
beenprofitablein a commercial sense.Theyhaveto beunderstood
within the more "traditional" framework of the imperial tribute
system. The motives were never commercial and it would be a
mistake to see them as an abortive attempt to establish a com-
mercial world-system, in what would become the European style.
They were intended to announce the might of the Ming emperor
'"3 See Dreyer 2007 for the history of the these fleets.
53
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
andbring foreign princes to recognise his suzerainty by rendering
tribute. 104
Such an interpretation would also help to make sense of the
situation prevailing in the following centuries, where the ban on
foreign trade was upheld for long periods. China is a large area to
control, and it turned out to be impossible to prevent even large
numbers of Chinese merchants setting out to sea. However, the
ban meant that the merchants were, so to speak, outside the sphere
recognisedby the state. Oneconsequencewas,thatthey oftenhad
to depend on the protection of local Chinese gentry, who knew the
"value" of their service; another, that they were left in a vulnera-
ble position vis-a-vis the emerging aggressive European merchant
empires. Thus we have examples of massacres in the Chinese
overseas trading communities carried out by various European
nations without the Chinese imperial state paying much notice. It
did not recognise the existence of these traders. In other words, the
interests of the commercial groups in Chinese society were never
articulated to the same extent as those behind European mercan-
tilist policies. 105 Therefore we do not find attempts by the state to
support the formation of commercial institutions as powerful as
those developed in cooperation by Europeanmerchants and states
to strengthen and stabilise the market system - state bonds, joint
stock companies and the like. It follows that we should imagine the
ChineseEmpire as going along quite anotherhistorical trajectory.
It was not a tailed or trapped move towards what we have come to
understand as modernisation. 10
Montesquieu more than 250 years ago provided an important
clueto the understandingof the reasonsofthis development, when
he seizedon the difference in state formation processes. 107 Focus-
ing his attention, as did many Enlightenment intellectuals, on the
large Asian empires facing Europe and its traders, such as the
Ottoman Empire, SafavidPersia,Mughal IndiaandCh'ing China,
hewasstruck bythecontrasts andnotedthatlargeagrarianempires
104Pearson 1991: 104-105; Landes 1998:93-98. EIvin 1973: 220 describes the voyages
as "prestige ventures". In general, see the analysis ofDreyer 2007: chapters I and 3.
105 Elvin 1973: 215-225: Gungwu 1990.
" Wong 1997:chapters4-6 is a pertinentdemonstrationofthis. Pomeranz2000underes-
timates issues of state formation and institutions, cf. Bayly 2000 and Pamuk 2000.
107 See Roberto 2003 for an interesting analysis ot'Montesquieu.
54
BEYOND THE ANCIENT ECONOMY'.'
had always failed in Europe whereas they seemed to prosper in
Asia. The reasons for this "failure" are now sought in the feudal
society that arose in Europe after the disintegration of the west-
ern parts of the Roman Empire. 10 An important characteristicof
this social system was the fragile position of centralised forms of
power, which produced a very fragmented kind of social forma-
tion split up into many small units. There were several attempts
to re-establish a strong imperial system in the European theatre.
The dramatis personae include such prominent figures as Charle-
magne, the Catholic Church and the Holy Roman Empire. Yet,
these ambitions always faltered, if sometimes only just, and suf-
fered their final defeat with the compromise in Augsburg in 1555.
This settlement left in tatters whatever was left of the dreams of
the Habsburg emperor Charles V to unite most of Europe under his
authority. Disillusioned, the emperor left it to his brother Ferdinand
to negotiate the peace and instead began to organise his abdica-
tion in the following year. This saw the division of the Habsburg
heritage into an Austrian and a Spanish branch.
Instead of an empire, a system of smaller competing states grad-
ually evolved out of the feudal world. Yet, central royal power
remained relatively weak against the multitude of feudal baronies.
For our purposes, the most important result of this state of affairs
was that the king often had to ally with mercantile groups to con-
solidate his authority. This brought these groups into much closer
proximity to state power than had been the case historically. To
begin with, this gave rise to the numerous strong medieval cities
where the merchants were allowed freedom of rule to a very large
degree. Sometimes they even became the government, such as in
the rich communes of northern Italy, with Venice, Florence, Pisa
and Genoa leading the way, and in the northern parts of Europe with
the Hanseatic league. Trade and the interests of commerce were
fast becoming of great importance to the state. As arms technology
developed and became much more expensive, especially with the
Landes 1998: chapter 3, though too simplistic. For the general processes at work see
North and Thomas 1973; Anderson 1974; Wallerstein 1974: chapters 3-4; Braudel
1981; Jones 1981; Braudel 1982: McNeill 1982; chapters 3-4; Braudel [984; Wickham
1985; Mann 1986: chapters 12-15; Burke 1988; Jones 1988; Kennedy 1988: chapter2;
Crone [989; Mauro 1990; Brady 1991; 0stergaard 1996; Bayly 2004.
55
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
introduction of artillery andnew forms offortification designed to
withstand bombardment from cannons, royal power was strength-
ened in relation to the feudal barons, who could not afford the
new technology on their own. Centralisation followed and the
nation states ofEurope came into being. However, increased inter-
state competition and the enormous, continuously rising costs of
warfare caused by the ensuing arms race made it crucial for state
powerto cooperate withallpropertied groupsofsociety inorderto
secure ever larger revenues. Thus the alliance with the commercial
sections of society wascontinued; numerous measures, known in
the convenient shorthand as mercantilism, were introduced, even
aggressive warfare, in order to favour andpromote the interests of
the traders of each nation, with the motive of generating as much
income aspossible. Themilitary balanceofpowernow wenthand
in hand with an unprecedented kind ofeconomic interstate rivalry.
The result was a much strengthened and stabilised institutional
base for the mercantile sector of society. '09
Characteristically, the states which formed the closest alliance
with the commercial parts of society proved strongest during the
seventeenth and eighteenth centuries. Hence the spectacular rise
of the Dutch Netherlands and the repeated British victories over
much larger France. On the face of it, one would have expected it
to be otherwise, given the fact of French absolutism and the need
for the weaker British state/king to come to terms with parliament.
Voltairecapturedthe situationvividly whenhe noted that
WhathasmadeEnglandpowerfulis thefactthatfromthetimeofElizabeth,all
parties have agreedon thenecessity offavoring commerce. The sameparliament
thathadthekingbeheadedwasbusywithoverseastradingpostsasthoughnothing
washappening.ThebloodofCharlesI wasstill steamingwhenthisparliament,
composed almost entirely offanatics, passedthe Navigation Act of 1650. "°
Interestingly, amongthe statescontendingfor a dominantposi-
tion in Europe, it was Habsburg Spain - the state formation most
resembling an old-style territorial empire - in spite of all her gold
and silver flowing in from the New World, that proved least able
to accommodate the demands of the new situation of intensified
iw) Hont 2005. De Vries 1976 is excellent on the resulting institutional arrangements.
"° Cited from Landes 1998: 234. See also Gellner l988b: 114-115.
56
BEYOND THE ANCIENT ECONOMY?
commercial rivalry. It was, quite simply, difficult to run a vast
territorial empire along the same lines of policy.
An early and almost too perfect illustration of the difference
between agrarian empires and the emerging new organisational
forms in Europe is provided by the relationship between a num-
ber of Italian merchant republics, with Venice in the forefront,
and the surviving, Byzantine, part of the Roman Empire in the
first centuries of the second millennium. During a period ofpolit-
ical and military weakness the Byzantine emperor found himself
forced to concede various trading privileges to the Italian mer-
chants in return for military support. In effect, a situation arose
where the Empire's own merchants were put at a disadvantage in
Constantinople in comparison to the Italian merchants - this was
mercantilist protectionism stood on its head. The point about this
is not its "perversity" from a European point of view, as Donald
Treadgold, the author of a recent synthetic history of the Byzan-
tine Empire, would have us believe. It is rather that the economic
weight of the commercial groups was normally simply not suffi-
ciently important in a large agrarian empire, which could draw on
income from huge masses of peasants, ever to give rise by itself to
a similar strong position for the mercantile sections of society.
This does not mean that merchants in such imperial formations
were entirely powerless. We can turn to the richest trading city of
the Mughal Empire, Surat, in the eighteenth century for indica-
tions of the limits within which they could act. Rising up against
an unusually extortionate governor, the rich merchants of the city
actually managed to drive him out of the town. But that was the
end of it. There was no attempt at systemic change, and they had
to settle for a new governor, belonging to the nobility. Thus the old
order was restored. "3
All this makes early modern Europe a very bad guide indeed
for understanding the economies of the large agrarian empires of
Eurasian history. Quite simply, they do not appear to provide any
''' Wallerstein >974a: chapter 4; De Vries 1976: 249-250; North 1990: 112-1 15.
''2 Treadgold 1997: chapters 18-21, especially pp. 699-706. On p. 704 the situation is
described as "the perverse policy of reverse protectionism". Pearson 1991 and Wong
1997: chapter 6 make the point in general for these empires.
' '3 Das Gupta 1979: Chapter 4.
57
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
strong parallel to the pattern of institutional developments we find
in the competitive state system of the early modern period. The,
at times, very large observable amounts (hence the modernising
critiques of "othering" analyses) of trade in these empires were
located in a different structural and institutional setting. In some
respects, the empires even seem incompatible with the expanding
capitalism of the eighteenth and nineteenth centuries. There is a
strong argument that they may have fallen victim to its develop-
ment. A large and growing literature focuses on how the Ottoman
Empire found it very difficult to keep up with European develop-
merits and was gradually undermined and incorporated into the
European economic system. "4 The Mughal Empire had disin-
tegrated before the beginning of the British conquest of India,
though there were perhaps signs that a reestablishment of imperial
authority wasimminent under the leadership of one or other ofthe
Mughal successor states. In the event, it all came to nothing with
the rise of British supremacy. More important from our point of
view is thepossibility that theinitial weakening ofimperial author-
ity came about through the increased involvement of India in the
growing European worid-system. The resulting large imports of
bullion into the Indian economy as payment for spices and cot-
ton calicoesmay havecontributedto underminingthe established
power structure. "5 The humiliating defeat of the Chinese Empire
in the Opium War (1839-42) spelled out to the authorities the need
for modernisation and industrialisation. But this proved a goal dif-
ficult to achieve within the already existing system. Eventually
"industrialization did take place in twentieth-century China, but
this China was no longer an agrarian empire. ""6
Thus it appears futile to insist on envisaging these empires in
analogy to the competitive system of more or less mercantilist
nation states arising in early modem Europe. Such comparisons,
whenscrutinisedclosely, seemdestinedtoendup in a catalogueof
failures and near misses. The large traditional or classic agrarian
'14 See Keyder 1997; Landes 1998; chapter 24; and Pamuk 2000 for brief surveys. Inalcik
and Quataert 1994: vol. 2, parts 3 and 4 for greater detail. Islamoglu-Inan 1987 and
Masters 1988 for world-system-inspired approaches.
115Steensgaard 199011.Wink 1989for re-establishmentofcentral control.
116Wong 1997: l5l. McNeill 1982: 257 aboutthe OpiumWar.
58
BEYOND THE ANCIENT ECONOMY?
empires were, for all we know, not on the same path of historical
development. The strategy of reverting to explanations of the func-
tional equivalent type is ultimately unsatisfying, as I tried to argue
from the example of Picket. Being mainly defensive in character,
it only succeeds in taking a very narrow view of the phenomena
and ignoring their markedly different contexts. For the study of the
Roman imperial economy this carries, at least, two implications.
First, it is time, finally, to abandon the tyranny of Europe over
Rome. The unusual pattern of European history cannot do service
as a yardstick to measure the very different imperial experience.
Second, in terms of comparative history it seems reasonable to
suggest that analogous or parallel comparisons to help understand
the institutional and structural characteristicsof the Empire should
rather be sought among the Asian empires that so impressed the
Enlightenment intellectuals.
Certainly, in terms of sheer scale, the Asian empires constitute
a much more fitting parallel to the Roman world. From a histori-
cal perspective, all of them, like Rome, are characterised by quite
substantial amounts of trade and urbanisation. They also seem to
show significant similarities in the structural location of commer-
cial groups and the market sector in society. In respect to state for-
mation, they represent an alternative path of development during
the early modern period to that experiencedby Europe. While the
feudal fragmentation of Europe eventually crystallised into a sys-
tem characterisedby medium-sized states, these empires emerged
as hegemonic powers dominatingtheir regions. The establishment
of the Ming dynasty in 1368, later succeded by the Ch'ing in
1648, saw the restoration of a Chinese Empire dominating East
Asia. During the fifteenth and sixteenth centuries, the Ottoman
sultans conquered vast stretches of the Mediterranean and Mid-
die East. Finally, the Mughals rose to dominate India during the
sixteenth and seventeenth centuries. None of these empires was
initially locked into interstate competition to the same extent as the
emerging nation states of Europe. Therefore they were not under
the same pressure to generate new revenues from their existing
tax bases to survive, for instance by granting commercial privi-
leges. Instead they enjoyed the benefits derived from expansion.
Having conquered many of their neighbours, each of the empires
59
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
controlled a much bigger agrarian economy than the European
monarchies. The imperial governments were able to field vast
armies by relying on collecting (land) taxes from their extensive
provincialterritories; they were, in that sense,"tributaryempires"
rather than mercantilist states. 117 Enjoying the benefit of at least
another millennium of slowly accumulating social developments,
these empires hold the further attraction as comparandathat they
are unlikely to be more "archaic" than Rome in any meaningful
sense. Making a shift in our comparisons of the Roman world to
this group of Asian tributary states, therefore, would allow us to
explore in greater detail the relationship between market system
and agrarian empire, in other words to understand it as a particular
type of political economy or system. The next chapterwill begin
to explore how this can be done.
117Woolf1990points in thesamedirection. Pleket 1990: 151remainsscepticalon thedis-
tinction between empire and mercantilist state system, but does not see the consequences
of the very different articulations of commercial interests.
6o
AN AGRARIAN EMPIRE BETWEEN MARKET
AND TRIBUTE: SITUATING INTERREGIONAL
TRADE IN THE ROMAN WORLD
[Tjhe decline of Rome was the natural and inevitable effect of immoderate
greatness ... as soon as time or accident had removed the artificial supports, the
stupendous fabric yielded to the pressure of its own weight. The story of its
ruin is simple and obvious; and instead of inquiring why the Roman empire was
destroyed, we should rather be surprised that it had subsisted so long.
Gibbon, The Decline and Fall of the Roman Empire, vol. IV, p. 119
Those are the immortal words of Gibbon, uttered with his charac-
teristic acrimony while contemplating the decline and fall of the
Roman Empire. At first, they may come as a surprise. After all, this
was the very same man who proclaimed the age of the Antonine
emperors "the period in the history of the world during which the
condition of the human race was most happy and prosperous".1
But to Gibbon, the golden second century was an aberration. Only
a fortuitous succession of unusually capable, conscientious and
moderate emperors had made it possible. In the long mn, the cor-
rupting influence of absolute power was bound to prevail and turn
the enlightened monarchy into an oppressive tyranny. Ironically it
was left to Commodus, son of the philosopher king Marcus Aure-
lius, to bring out the true face of imperial rule. With his accession
to the throne the history of the Empire became one of despotism,
corruption, military anarchy, brutal oppression and the withering
away of its powers. And yet, in spite of its alleged pernicious influ-
ence on society, the Roman Empire lasted for centuries. Indeed,
it came as a shock to Gibbon in writing his history to realise the
almost, from this point of view, perverse ability of the imperial
system to renew itself andregain its strength;2 he needed six long
volumes finally to lay the leviathan to rest.
' Gibbon 1993, vol. i : 90 and further 90-99. Gamsey and Humfress 2001, chapter 10 on
Gibbon.
2 Robertson 1997.
6i
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
The relationship between the predatory and benign aspects of
imperialrule remainsa key issuein ourunderstandingoftheeffect
ofRome andotheragrarianempiresontheireconomies.Thischap-
ter argues that these two aspects of the imperial experience cannot
be separated, but have to be seen as two sides of the same coin.
Economicdevelopment wasthe result ofpolitical exploitation, not
its victim. This requires us to shift our attention from markets
as the organising force in the economy to the process of trib-
ute extraction and mobilisation of peasant surplus production. A
comparison of the interlinked processes of surplus extraction and
elite building in the Mughal and Roman Empires demonstrates
the importance of imperial tribute in determining economic flows
in the empires. Markets, from that perspective, do not determine
the structure of the economy; instead they play a crucial role in
mobilising the agricultural surplus and turning it into a disposable
resource. The chapter concludes by discussing to what extent the
processes of tribute extraction spawned economic growth in the
Roman Empire.
Agrarian empires between market and tribute
Ultimately, the universal empire had presented itself as a paradox
to Gibbon, a perception he shared with many of his contempo-
raries. In The Wealth of Nations, Adam Smith described China of
the Ch'ing dynasty as one of the richest societies in the world with
a highly productive irrigation agriculture, substantive division of
labour and an extensive inland trade. But at the same time he saw
the rule of the "Son of Heaven" as inimical to the economy; the
imperial tributary elite depressed the economic performance of
the country through political privileges, taxation, arbitrary admin-
istration andoutrightpredatory encroachmenton the wealthofthe
producing and mercantile layers of society. 3 Market and tribute
existed in contradiction to each other; they seemed incompatible
and yet they were co-existing. The Enlightenment philosophers
never really managedto dissolvethis tension andincorporateboth
3 Smith 1976: 11 i-l i2.
62
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
elements in their understanding of the great empires. Some, like
Voltaire, chose to ignore the "darker" side and instead eulogised
the imperial claims to absolute power. In his Essai sur les moeurs et
I 'esprit des nations he presented imperial China as the paragon of
rational order, an idealised counter-image to ancien regime France.
This solution tailed to win the majority opinion. Most intellec-
tuals were more impressed by the oppressive effects of imperial
rule. In L'Esprit des lois Montesquieu fiercely denounced the rosy
image of China; tributary empires were based on the principle of
fear. 4 Later, Adam Ferguson followed suit in his tract on civil soci-
ety. The absence in the empires of a close alliance with commercial
groups and the concomitant predominance of political consump-
tion (cf. Chapter i), was fundamentally at odds with the world
view of the increasingly powerful and confident bourgeois soci-
eties of Europe. Empires seemed the embodiment of everything
adverse to the kind of social order they promoted, with clearly
defined civil liberties and parliaments to contain the ruling power.
In the bourgeois view, the empires appeared to be antagonistically
opposed to their very economic and societal bases. Here is one of
Adam Ferguson's damning observations:
Were despotic empire... to continue for ever unmolested from abroad ... it
appearsto have in itselfno principle ofnew life, and presents no hope ofrestora-
tion to freedom and political vigour. That which the despotical master has sown,
cannot quicken unless it die; it must languish and expire by the effect of its own
abuse... National poverty, however, and the suppression of commerce, are the
means by which despotism comes to accomplish its own destruction.5
Stagnationand misery were all that agrarianempires had in store
for their poverty-stricken subjects. Commerce and the market were
left very little room in the imperial order.
It took Max Weber to realise that the picture was much over-
drawn and strongly stereotyped; empires would not have been able
to exist under such purely traditional conditions. Here he made the
fundamental observation that once a traditional, or in his termi-
nology, patrimonial state began to grow, the household of the ruler
would no longer be able to control the entire area of his realm.
The ruler had to form alliances with local communities, mainly
Montesquieu 1955, part I: 3, 9-10; 7, 7 and 8, 21. 5 Ferguson 1995: 262-263.
63
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
aristocracies, in order to establish longer-lasting control and have
representatives carrying out his commands. The much dreaded
despotism of the sociological tradition was not really possible.
The power of emperors and the imperial state was only appar-
ently absolute and arbitrary. In practice, the ruler over extensive
areas had to negotiate and accept limitations on his exercise of
power. Consequently the empires were characterised by a tension
or uneasy co-existence between the tax-extracting central power of
the emperor (tribute) and locally entrenched power wielders (feu-
dal elites), living off agricultural rents. The resulting institutionali-
sationofgovernmentinto relatively stablepatterns of activity was,
asWebersawit, thebeginningsofmore complextypes of society
with a more pronounced specialisation of functions. This led him
to include his treatment oftributary empires in a general discussion
ofthedevelopment ofmodern forms ofstateandbureaucracy.6
For our present purposes, however, this choice has had some
unfortunate consequences. An opportunity was partly missed to
break free from the contradictions in the Enlightenmenttradition.
In Weber's analysis of the origins of modem states, the crucial
factor became the ability ofthe feudal level ofsociety to curb what
seemedthearbitrary aspirations ofthetributary state. However, the
more successfully this was done, the less chance the empire stood
of survival. Only where feudalism had been triumphant and the
imperial superstructure had disintegrated, did society set off on a
developmental path eventually leading to modernisation. Inthe last
analysis, the antithetical relationship between empire andits econ-
omy and civil society therefore remained central for Weber. The
tributary systems may have shown some signs of institutionalising
social and economic diversification. But ultimately the two phe-
nomena were incompatible. Conceptually, this resulted in a hybrid
characterisationof these systems aspatrimonial-bureaucratic.7
That was problematic. It was the tensions in Weber's concep-
tualisation which came to define the path ahead for later work
in historical sociology. Several influential studies in the Webe-
rian tradition have tended to disregard the cautiously balanced
6 Weber 1972:580-652. See also Wickham 1985.
7 SeeWeber1972:619.Seepp.593,604and625-653forhiscontinuedadherenceto the
Enlightenment interpretation.
64
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
observations of Weber and instead accentuated and aggravated the
internal contradictions in the concept of tributary empire. In The
Political System of Empires (1963) Shmuel Eisenstadt attempted
to develop Weber's basic analysis of complex agrarian societies.
Eisenstadt's analysis is not without merits. Its focus on the ability
of the empires to generate "free floating resources", that is their
capacity for mobilising the agricultural surplus outside its imme-
diate context, has much in common with the view which will be
presented here. But the study has one crucial drawback. It treats
the empires entirely within a framework of very abstract 1950$
modernisation theory. This prevents the empires from appearing
in their own right. Analysing Roman imperial administration in
terms of a modem-style bureaucracy is, for instance, not very
revealing when we remember that it had no institutions which
served to guaranteethat promotion happenedon the basisofmerit
or professionalqualificationratherthanthroughpersonal favour.8
Eisenstadt s approach tends, in other words, to miss some of the
important organisational characteristics of the empires. They are
perceived as incipient cases of modernisation still needing to shed
much of the dead weight of traditional society before coming
to fruition. Essentially this confronts agrarian empires with two
options: eitherthey transform successfully into mercantilist nation
states and modernise or they remain traditional empires and even-
tually kill off whatevermodern elements they possess.9
John Kautsky has taken this further in The Politics of Aris-
tocratic Empires (1982) by sharpening the contrasts appearing
in Eisenstadt s image. Here the empires are simply defined as
essentially traditional. Trade and other such phenomena are, on
the other hand, seen as signs of modernisation per se and hence as
fundamentally outside the aristocratic order. This is a very unfortu-
nate solution. Most agrarianempires andcertainly the best known
had considerable amounts of trade taking place within them. But
Kautsky'ssystemcannotreally accountforthis. Commercialphe-
nomenaare left in a kind of analytical no-man's land betweenthe
aristocratic and the present world until the agrarian or tributary
empires finally hadmodernisationforced uponthem from without
Saller 1982: 111-116. 9 Eisenstadt 1963:351-360.
65
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
by a conquering European modernity and in that process were
transformed into a new type of society. 10
A conceptualisation oftrade andeconomic activity that situates
these phenomena firmly within the tributary empires constantly
eludes us. The relationship between traditional empire and trade
seems todefy ouraccustomed categories. According toestablished
notions they ought to be like oil and water: mutually exclusive.
The essence of the orthodox view is neatly summarised in John
Hall's notion of capstone government. Apparently, an imperial
centre could only maintain control by weakening its underlying
society and economy. For its own survival, it was forced to prey
on any dynamic development and finally to block and destroy it.
In that way, the empire prevented local groups from becoming
wealthy and strong enough successfully to challenge its authority
and refuse to pay taxes. Thus imperial rule was secured by creating
a power stand-off whichlowered the level of social activity and
put a lid on creative energies. Empires had"strong blocking but
weak enabling powers", as Hall has phrased it."
Redistributionor integratedmarket systems?
To be sure, it is not difficultto findexamples of imperial authori-
ties abusing their power to plunder and depress the subject popu-
lations. A recent book on Roman Syria stresses how the imperial
army damaged the local economies to a large extent. The sol-
diers often proved a dangerous presence, being only too prone to
terrorise provincial inhabitants in order to extort money or other
desirableresourcesfortheirownbenefit. '2 Yet,therearesomevery
important aspects ofthe imperial experience for whichthewhole
label of capstone government seems inappropriate. For instance,
the empires were able to create some of the largest urban concen-
trations in the pre-industrial world. Metropoleis such as Mughal
Delhi, Peking and Rome represent, by pre-modem standards, not
Kautsky 1982, chapters l, 2 and 15.
Hall 1985: 35.Seefurtherchapter2 andpp.99-109.Mann 1986,chapter5 fora critique
of this whole tradition.
; Pollard 2000, chapters 3 and 7.
66
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
only an enormous concentration of spending capacity, but also
with populations reaching up towards the million mark (or above),
a gigantic logistical accomplishment. 3 "Low enabling powers'
does not appearan adequateexpression with whichto capture such
phenomena. Thinkers less suspicious of state power have there-
fore been more inclined to note a paternalistic, beneficial side to
imperial government. This point of view can be traced back to the
Jesuit missionaries reporting about the enlightened, orderly rule
of the Chinese emperor to the reading public of Europe in the sev-
enteenth and eighteenth centuries. 4 In more recent times, Weber
tentatively suggested that the imperial inclination to give priority
to consumer over producer interests was a first premonition of the
welfare state. 5 More than anyone else, however, it is Karl Polanyi
who has given expression to this view with his notions of redis-
tribution and administered trade. Disgusted by capitalism he set
about exploring historical alternatives to a modem market econ-
omy. One such alternative he found in a group of ancient Middle
Eastern empires. Apparently the economies were organised with-
out price-fixing markets. Instead resources were allocated within
a framework of state-controlled redistribution of production and
legally fixed prices. To Polanyi's mind, this was a planned econ-
omy of a sort and he did not hesitate to draw parallels between
his ancient redistributive empires and the economy of the Soviet
Union; essentially they wereof the samekind.
Polanyi's ideas have proved fertile in Roman soil. They have
stimulated a substantial number of studies which have been able
to show the large influence that state-administered redistribu-
don exercised on the movement of material resources around
the Roman Empire. The archaeological record showing the
distribution of exported oil and wine amphorae has, for instance,
3 See Braudel 1981: 489-491, 498-504 and 525-559 about pre-industrial metropoleis;
Blake [991 on Delhi; Morley 1996 on Rome.
14 See Bitterli 1991: 60-65 and 270-272 about the place of China in early modem Euro-
pean discourse. The reports of the Jesuits had a strong influence on philosophers such as
LeibnizandVoltaire. Oneofthecentral works in thegenreis fatherDu Halde's Descrip-
tion geographique, historique, chronologique, politique et physique de la Chine et de
la Tartarie, Paris 1735. Whereas Voltaire was delighted with it, Montesquieu, with his
doctrine of the division of power, attacked it vehemently in L'Esprit de lois (1955, part
1:8, 21).
15 Weber 1972:651-653. '<> Polanyi,ArensbergandPearson 1957,chapters 2 and 13.
67
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
to some extent been persuasively linked to the so-called annona
system, or rather systems which the state used to organise food
supplies for Rome and the armies. 17 Yet, in its pure form the
notion of redistribution cannot be elevated to the organising prin-
ciple of the imperial economy. It is in a way, modernising in spite
of itself. Polanyi's redistribution simply exaggerates the organi-
sational strength of imperial systems such as that of the Romans.
From a modern perspective the striking thing is rather the very
limited capacity for bureaucraticcontrol of pre-modem states and
often their outright impotence.
Some administrative documents from the Egyptian nome (dis-
trict) Panopolis, dating to around AD 300, offer ample illustration
and deserve to be presented in some detail. Two papyrus scrolls
contain copies of letters sent to and from the office of the strate-
gos, the local representative of the imperial authorities. 19 From this
correspondence we learn that the strategos was expected to sub-
mit monthly accounts of the collection of taxes and the expenses
incurred in government service. However, the predecessor had
neglected his duties for some time and had now left the area without
leaving a proper record behind. In practice, therefore, the accounts,
which according to imperial command and expectation ought to
have been kept in scrupulous order, were in a mess and could not
easily be corrected. 20 Ofcourse, we know ofthis affairbecausean
attempt was made to remedy past neglect, though precisely with
what effect is uncertain. The letter scrolls do not uniformly docu-
ment governmental failure; they also reveal an administration, at
least periodically, attempting to tackle attempts at evasion. But,
they do indicate that the gap between imperial prescription and
reality would often have been a wide one. The scrolls are also full
of letters from higher-level officials giving and repeating orders,
17 Peacock and Williams 1986, chapter 5. Some single studies emphasisingstate redistri-
bution: Midclleton 1983; Wickham 1988; Whittaker l994a; Foraboschi 1994; Remesal
Rodnguez 1997.
18 Saller 1982,chapters 3 and5; GamseyandSaller 1987,chapter2.
19 P. Beatty Panop. I and 2.
20 P. Beatty Panop. [, lines 64-71 and 90-107. The problem of disorderly or imperfectly
submitted accounts occurs repeatedly in the correspondence, e. g., P. Beatty Panop. i,
lines 72-76; 2: lines 10-14 and 61-67.
21 Cf. P. Beatty Panop. 2, lines 68-71 and 128-152(seeking a pragmatic settlement rather
than effectuating punishments for tax evasion).
68
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
often several times over, that had so far been disregarded locally
in spite ofdire threats ofheavy andfrighteningpunishments. Even
officials in chargeofthe cultivation ofgovernment landcould not,
as a rule, be expectedto obey orders promptly, if at all. 22
Thus, on closer inspection the idea that the imperial economy
was controlled by a large redistributive system that organised all
or most of the transfer of resources in the Roman world has shown
itselfto beinadequate andalso somewhat unrealistic. 23The Empire
simply lackedthebureaucraticcapacityfor sucha thing.Theredis-
tributive measures were very restricted in number and scope and
only comprised a small part of the economy. In the first two cen-
turies AD they weremainly limitedto supplyingmuch, thoughnot
all, of the grain needed to feed the population of the capital, and
most, though even that is contested by some, of the requirements
of the army. 24 There are indications in the sources that the system
grew larger, mainly from the Severan dynasty onwards. But the
extent is quite unclear and it has recently been concludedthat the
evidencedoesnotlendsupportto anynotionofradicalchange.The
system remained limited in scope.25 Correspondingly, the pendu-
lum is now swingingbacktowards (re-)assigningthe central state
apparatus a traditionally more marginal position in the economic
life of the Roman world.
As was to be expected, while the dimensions of the state have
beenshrinkingin the eyes ofmoderncommentators, so the impor-
tance of private market trade has received new and increased
emphasis, as already noted in Chapter I. One theme which in
particular has been able to command scholarly attention in the
last decade or more is the possible formation of large markets and
their integration on an empire-wide scale.26 Much of this reads
like a catalogueof "free trade" promotion. The role of the Roman
" For repetition followed by threats, see for instance P. Beatty Panop. I: lines I 10-119
and [66-179 (lines 184-187 show one of the repeated orders having gone through).
For officials on state land ignoring orders P. Beatty Panop. 1, lines 205-212. Another
example of "juicy" threats P. Panop. l, lines 100-108.
3 Harris 1993:14-18.
24 Garnsey and Saller 1987, chapter 5. Harris 1993: l-j is too sceptical about state control
of military supplies.
5 Gamsey and Whittaker 1998.
26 Fulforci 1987; Carandini 1988; von Freyberg 1989; Harris 1993;Parker 1996; Harris
2000.
69
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
state is mainly perceived as indirect, assisting the trading world
more or less consciously by providing a better infrastructure. In
this context many old ideas have been recycled and brought to
new prominence. Here we find the notion that the Fax Romana,
the imposition of a common currency, standardisedmeasures, the
highlydeveloped system ofRoman law andmore orderly adminis-
tration, all workedto improve economic efficiency, lowerthe costs
oftradingandpromote market integration. CharacteristicallyHop-
kins' model, linking imperial monetary taxationwithinterregional
trade flows, which was discussed in the previous chapter, has been
adopted within this picture, not as providing the dominant mech-
anism, but at most as just one factor within the whole range of
economic catalysts. 27
Somego even furtherandrelegate the state to a position ofpure
exploitation on the fringes of the economy. In The Corrupting Sea
Peregrine Harden and Nicholas Purcell argue that the pre-modern
states in the Mediterranean mainly worked to intensify exploita-
tion, sometimes with locally disruptive effects.2 The basic pat-
tern in the economy, however, was defined rather by the regional
ecology. Accordingto their argument, extreme geographical frag-
mentation, splintering even micro-regions into many sub-types,
combined with an unstable and erratic pattern of annual rainfall to
make subsistence agriculture an impossibility; there was no such
thing as an average harvest which the peasants could depend upon,
only extreme variability. The much praisedandidolisedautarkyof
the peasants is a modem myth. In the Mediterraneanreality, aim-
ing for self-sufficiency was a suicidal strategy. To survive, peasants
had to rely on surplus production and exchange. "Every crop is a
cash-crop", the authors assert. The crucial role of exchange was
muchhelpedby the seawhichservedto bindthefragmentedlocal-
ities together. It made the transfer and exchange of agricultural
products between individual localities a fundamental feature of
the Mediterranean world. Interdependence and economic integra-
tion, thoughvarying with levels ofexploitation, were, so to speak,
27 Lo Cascio zoooc is the best-argued example. Paterson 1998 is extreme, rejecting the
relevance ot'Hopkins' model.
28 HardenandPurcell 2000,chapter7.
70
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
written into the DNA of Mediterranean civilisation. Here is one
of their observations: "We may even regard the state of affairs
today, in which the Mediterranean lands produce a fifth of the
planet's vegetables and a third of its fruit but depend on imports
for two-thirds of their own grain, as a logical extension of the
intensification that we have been discussing."29 Of course, there
is an element of rhetorical hyperbole here which perhaps ought
not to be given too much weight. Yet, William Harris has made a
comment to much the sameeffect, namely that the economic inte-
gration of the Empire was probably equal to that of the European
market system including the colonies around l750. 30
These are extravagant claims, as can be briefly shown with the
aid of a few examples. Let us leave the position of the Mediter-
ranean in the modem world economy aside andtake a closer look
at the formation of the early modern European colonial market
system. During the Middle Ages the sugar consumed by Europe
was produced in the Mediterranean area. But that was almost
brought to a complete end when sugar cultivation was moved, first
to the plantations of the Atlantic islands and then to Brazil and the
West Indies, whose production systems became completely organ-
isedaroundthis activity. Laterthe samethinghappenedto coffee,
which was moved from Yemen to Java and the plantations of the
New World; meanwhile tea was brought from Chinato India and
Ceylon. The drivingforcebehindtheserelocations oftheproduc-
tion of export crops was the search for greater profits; in the new
locations production was either cheaper, more easily controlled
and organised or had greater room for expansion. 31 Sugar, coffee
t! HardenandPurcell 2000: 279 and 273 for the first quotation. Pp. 270-274 summarise
theirviews. Chapters 5, 6 and9 develop theirview of Mediterraneaninterdependence,
connectivity in their terminology. The Corrupting Sea is a bold work of synthesis and
should be admired for that. However, it is also a very difficult book and does not
always appearconsistent in its views. For instance, while elite exploitation is seen as
anextensionofthebasicproductivepattern, it is still mostly viewedasdisruptiveand a
threatto thewell-beingofthe"victim"population(278-287,incontrastwithp. 223),in
otherwords,asrepresentinga breakfromtheecologicalstandardpattern.Ifthatis so,it
becomescrucial to explorehowdifferentregimesofelite exploitationmouldeconomic
activity. However,that is precisely whatthey rule out by blurring distinctions between
the BronzeAge,classicalantiquityandtheMiddleAges. Fora critique,seeBang2004.
iu Harris 1993: 19;Temin 2ooi. 3I Steensgaard 19900and i99ob.
71
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
andtea, however, are only examples of a general process in which
the economic geography of Europe and large tracts of the world
were gradually reorganised to fit the needs of a commercial system
basedon a steadilydeepeningregionalspecialisation.3 Theunder-
lying principle received its classic expression in David Ricardo,
the father of modern international trade theory, when he observed
that foreign commerce and the resulting international division of
labour ensured that:
each country naturally devotes its capital and labour to such employments as are
most beneficial to each . .. while by increasing the general mass of productions,
it diffuses general benefit, and binds together by one common tie of interest and
intercourse the universal society of nations throughout the civilized world. It is
this principle which determines that wine shall be made in France and Portugal,
that corn shall be grown in America and Poland, and that hardware and other
goods shall be manufactured in England.33
Subscribingto a Ricardianvision oftheEmpire,in otherwords,
involves positing a radical, artificial, commercial remoulding of
productionpatterns in the Roman world. The result would haveto
bea systemwhereprovincial economiesweretransformedto serve
the needs of a steadily deepening interregional division of labour
based on regional specialisation, economic competition and com-
parative advantages in a conglomeration of interdependent mar-
kets. Spelledout in suchdetail, few,perhaps,wouldbepreparedto
accept the implications of some of the optimistic assertions made
on behalf of the Roman and ancient Mediterranean economy. It
is doubtful, for instance, whether this image is in fact what Hor-
den and Purcell intended to create in their analysis. After all, the
ambition to integrate history and ecology can be read as an attempt
to create a completely new framework for our understanding of
exchange activities in the ancient world. Even the old Polanyian
notion of redistribution is developed and expanded, in The Cor-
rupting Sea,to covera broaderrangeoftypesofexchange.Clearly,
an alternative to the Ricardian notion of long-distance trade is
needed.
32 SeeWallerstein 1974andWolf1982for a world-systemdescriptionofthisdevelopment.
33 Ricardo 1996: 93.
72
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
Interregional trade and agricultural surplus areas
Now, obviously all interregional trade which achieves a degree of
regularity implies some kind of interdependence; a great city like
Rome would not have existed without grain imported from the
provinces. But we are still far from the kind of economic integra-
tion demandedbythemodernisingmodel. Taketwoofourarchae-
ologically best-recorded products: wine andoil. Thebeginning of
wineimportstoRomeinthelateRepublic/eariyEmpirefromGaul
and Spain, carried in the so-called Dressel 2-4 amphorae, did not
bringanendto thedelivery ofItalian supplies to thecapital. -34Sim-
ilarly, the enormous growth in the early Empire in production and
export of Spanish and, later, African olive oil did not extinguish
oleo-culture in other areas of the Roman world. 35 The extent of
regional specialisation, in other words, remained quite limited. A
recent archaeological survey of the city of Leptiminus in modern
Tunisia offers a strong additional confirmation. 36 The town seems
to have had a very substantial export of olive oil and fish sauce
from the second to fourth centuries AD, to judge by the amounts
of waste from amphorae production found on the site. On the other
hand, the city had no local supplies of iron and had to cover its
needs through imports. Yet, the archaeological survey documents
the existence of a local iron "industry"; substantial amounts of
debris from the smelting of iron ore was found on the site. The
most likely explanation is that ore was brought in as ballast in the
ships carrying offthe oil. Significantly, it was apparently not more
expedient to import the metal after it had been processed into bars,
let alone manufactured into finished products. Undeniably, this is
Tchernia 1986. Panella and Tchernia 1994. Whittaker 1985: 50-51 summarises the
amphorae material.
35 Keay *984' Mattingly T988a;andRemesalRodriguez 1998are good treatments of
AfricanandSpanisholive oil. In late antiquity but beginningat an earlierdate, olive
oil, produced in the numerous villages dotting the limestone massifofnorthern Syria,
has also been interpreted as an instanceof regionalspecialisationin the pioneering
surveyworkofTchalenko. However,re-examinationofthismaterial in the lastdecades
hasunderminedhis interpretation. The agriculturein the areawas"amixedregime
dominated byarablecultivation" anddidnotdependontheexportofoilforitsexistence,
soPollard2000:201-203and232-236,buildingonTate 1992.
36 Mattingly et al. 2001.
73

THE ROMAN EMPIRE AND COMPARATIVE STUDY
an example of regional interdependence, but also ofvery fragile
and limited specialisation.
Thisconclusion oughtnottobecontroversial. Ithasbeenrepeat-
edly observed that Greek and Roman states failed to take com-
mercial advantage of their colonies in the way that early modem
European states'attempted to in the organisation oUhe Atlantic
economy. Montesquieu must have been among the first to make
this observation, when he noted that ancient states did not treat
theircolonies asopportunities forincreasingeconomic specialisa-
tion and the division of labour. Incontrast to the mercantilist strate-
giespursuedbyhis contemporaries, the ancientsdidnot attemPt
to secure the continued commercial dependence of the colonies
on the metropolis through the granting of export and production
monopolies. 3'7Inspiteofcountless claims tothecontrary archae-
ology'has, by and large, confirmed this picture. Interestingly, it
was"Rostovtzeff, the founder of the concorrenw paradigm, who
firstsuppliedtheevidence.Withthecomingofthe"Romanpeace
hefoundit increasingly difficulttomakethepatternsproducedby
thearchaeological datafitthe Ricardianmodel. Ratherthangoing
theirownway^specialisinginparticulargoods,theprovincesbegan
to emulate the products ofItaly. 38
This is the problem known in the historiography as the (very
questionable)'''crisisofItaly".39InthesecondcenturyBCconsid-
erablequantities ofItalianexports startto appearintheMediter-
ranean,especiallyintheAegeanandGaul.Tojudgefromarchae-^
oTogicalfinds,theseexports seemprimarily to haveconsistedof
wYn°e (carried in Dressel I and Lamboglia 2 amphorae) and fine
pottery(theblack-glazedCampanian,followedbythered-glossed
Arretine terra sigillata) riding "piggy back" on the shipments of
wine. Inthefinafphase oftheRepublic withtheconquest ofGaul,
the consoUdation'of empire in the eastern Mediterranean andthe
coming ofAugustus, thelevel ofactivity dropped significantly. In
the West, for instance, the wine exports from Italy to Gaul expe-
rienced a steep fall and then petered off during the first century
AD. In the same period, Gaul gradually established a domestic
37 Montesquieu1955,part4,2i,21. 3s Rostovtzeft'1957,chapter^3-5.^
39 Carandini 1988is themostoutspokenproponentot the decline01 iiaiy iiicuiy.
74
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
production of the former imports of wine and pottery and even-
tually even began to send substantial consignments of wine to
Rome. 40
Thus through the lenses ofmodem trade theory, the inclusion of
Gaul intheEmpireparadoxicallyappearsto underminetheexisting
regional specialisation; economic integration seems to suffer and
the Empire enters upon a process of Entspezialisierung(declining
specialisation), to quote the recent work of a German economist. 41
Rather than the expected increased regional division of labour
where Gaul would have specialised in a different product (say,
ham or iron tools) it could use in exchangefor Italianwine,Roman
exports collapsed and the province became self-sufficient in goods
it had formerly imported.42 According to modern economic per-
ceptions this apparent decreasingdivision of labour ought to be a
sign of economic crisis and decline. And this, too, is what Ros-
tovtzeffandhis modern Italian followers have suggested. The fall
in Italian exports has been seen as an indicationof a growingeco-
nomic crisis, as first Italy and then the Empire as a whole entered
a downwardspiral ofdecliningdivisionof labouranddiminishing
economic integration. 43 This is unconvincing for several reasons.
First, howeveroneestimatestheconditionofItalyfrom the second
century AD onwards, a hotly debated issue, the peninsula seems
to have been in good economic health at least throughout the first
century AD and probably for much longer. The changes that do
occur are better described as transformations than decline.44
Second, to describe the incorporation of Gaul in the Empire
as a process of decreasing economic integration is to take too
narrow a view of the relationship between province and imperial
-I" Panella 1981; 1993. Panella and Tchernia 1994 survey the evidence.
41 Von Freyberg 1989: 151-152. Woolf 1992: 283-293 also sees integration culminating
in the first century BC.
42 Gren ] 941 also noted how imports from Asia Minor to the Balkans were never balanced
by a counterflow of specialised goods produced in the Balkans.
43 For this school of thought Giardina and Schiavone 1981 and Carandini 1988.
44 Tchemia 1986 is a devastatingcritique. See furtherPatterson 1987; Whittaker 19946;
Vera 1994; andGiardina 1997, chapter5. The development ofItaly is better explained
in terms of developments in patterns of landholding and political office among the
political elites. See Patterson 2006 for an attempt to synthesise such a view of Italy.
Sallares2002raisesthequestionofepidemicdiseasesforourunderstandingofimperial
Italy.
75

THE ROMAN EMPIRE AND COMPARATIVE STUDY
centre. It significantly overestimates the economic effects of the
Italianwine'exports. Theseexports werenot the result ofa deep
economic interdependence between the two regions. Onthe con^
trary, they seem to have been the outcome of a very restricted
set'of circumstances, mostly depending on a very low degree of
integration between the two areas. During the age of expansion
the demand for "barbarian" slaves and metal increased in Italy.
With contacts increasing between Roman and Celtic civilisations,
the Gaulic nobility acquired a taste for wine, a product of the
Mediterranean. In'return for this exotic and luxurious product,
Gaulic nobles seem to have provided Roman merchants with an
abundant supply ofmetals andslaves ofCeltic origin. After Gaul
was made a province this traffic became unacceptable and came
to a swift end. With the loss of these very particular conditions,
the~wine trade was doomed. 45 At the same time, however, social
and economic integration became more extensive and involved
greaterresources. Gaulmayhavestopped sendingmany slavesto
Rome. but it now had to pay a far more substantial imperial trib-
ute andsupply troops fortheRoman army. 46The imperial system
made greater'demands onthe economic resources ofpost- rather
than'pre-conquest Gaul. On top of this, the province entered a
phaseofpartialadoptionofRomaninstitutions material culture
andproductive strategies. Urbanisation increased andproduction
bothgrewandbecamemorediversified. 47Itsimplydoesnotmake
mudTsense to describe this process of inclusion and intensified
contact spanning a much broader spectrum ofactivities asoneof
decreasing economic integration andspecialisation. Social com-
plexitydidnotdecline,itincreased Itis,inotherwords,difficultto
explaintheexperienceofRomanimperialismwithinaRicardian
framework. Economic integration developed according todifferent
principles.
In fact, a better comparison than the strongly articulated com-
mercialspecialisation thatwefindintheEuropeancapitalistworld^
system,isthesituationintheIndianOceanworldbeforethearrival
45 Tcherma1983;1986:74-ioo,14^146and158-172,and2006. ^ ^ ^^ ^^
46 TacitusHist.iv. 26and71 portraystributeandarmyleviesasimportantta<
the Gallic rebellion in 69-70.
47 Woolt'2001 and 1998.
76
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
of the British Raj. 48 Chaudhuri summarises the basic conditions
of interregional trade in that part of the world: "But the compo-
sition of trade.. . was not mainly determined by the nature of
that specialisation [i. e. interregional division of labour]. .. The
most important consideration was the ability of a local economy
to create a surplus over and above subsistence demand and to
maintain this level. "49 The generation of an agricultural surplus
was a key factor in enablingthe development ofother commercial
and productive activities within a region. Most of what happens
in the Roman Mediterranean fits nicely into such a pattern. In the
westernpart the link betweencommercial prosperity andsustained
agricultural surplus production is pretty straightforward, with agri-
cultural products dominatingthe scene. It is the regions showing
considerable agricultural expansion such as the Guadalquivir val-
ley in Spain or much of Roman North Africa which turn up in
the evidence as commercially dominant. The study of the eastern
Mediterraneanis, at leastin terms ofarchaeology,less developed.
Still, whatwe know seemsconsonantwith this model. Alexandria,
for instance, praised by Dio Chrysostomus as the leading entrepot
intheEast,wasalsothe"capital"ofEgypt,themostfertileagricul-
tural province in the Empire.50 Ephesos, another emporium often
mentionedfor its prosperoustrade, wasone ofthemost important
cities in the agriculturally rich province of Asia.5'
For what it is worth, the close connection between the creation
ofan agricultural surplus andcommercialprosperity also appears
as a dominating feature of a late antique work of commercial
geography of the Roman world known as the Expositio Totius
Mundi. The authornotes, for example, that the cities ot'Phoenicia,
famed for their trade, all had an abundant production of grain, oil
and wine.52 Inevitably, in the short-term, such a trading system
4S Habib 1999: 39-62 and Moo.svi 2000: 345-346 underline that a closely integrated
market system did not develop in India before the construction of railways in the late
nineteenth century.
49 Chaudhuri 1985:l82.
50 Strabo 798. 13 and Dio Chrysostomus XXXII, 35-36. Dio's description ofAlexandria's
grandeurcharacteristically startsby noting its basis in the rich Egyptianhinterland.
51 Strabo 641, 24. Pleket 1994 and Kobes 1999 emphasise the importance of trade to
Ephesos,thoughthey underestimateagriculture, in my view.
5 Expositio Totius Mundi 23-33 (confirmed by MiIIar [993: 274-295). The close link
betweentrade andfertile cultivation is ubiquitousthroughoutthework.
77
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
would be susceptible to fluctuations causedby harvest failures or
other shocks to the agricultural economy that eroded the surplus
andthus"driedupsuppliesofcommercialgoodsanddestroyedthe
livelihoods of both artisans and farmers". To return to our Indian
comparison, during a disastrous famine in the province of Gujarat'
in 1630 the substantial export production of cloth dwindled away
as scores of spinners, weavers and peasants alike either fled the
province in search of food or succumbed to hunger. 53 A similar
image of the vulnerability of the commercial sector in the face
of agricultural failure in surplus areas is painted by a late Roman
chronicle describing a famine in Edessa: "Everything that was
not edible was cheap, such as clothes and householdutensils and
furniture, for these things were sold for a half or a third of their
value, and did not suffice for the maintenance of their owners,
because of the great dearth of bread."54
Catastrophicfamines,however, wererare. Normally the fluctu-
ations in output would be less severe, though far from insignifi-
cant.55 In the longerterm, otherfactorstake on greaterprominence
in shapingthe sizeanddistributionofthe surplus. Chaudhurilists
conditions such as the general fertility of the area, population den-
sity and the level oftechnology. One factor, though, is conspicuous
by its absence: surplus extraction. One of the great accomplish-
merits of the vast body of peasant studies has been to make clear
that peasanthouseholds are not normally orientated towards pro-
ducing a large surplus for circulation outside the farm. They aim
rather to fulfil the basic needs of the household with as little effort
as possible. Usually they have not been the main beneficiaries of
theproductionofsubstantialsurplusesabovetheirownsubsistence
needs. The bulk of their surplus produce was normally claimed by
various political lords and would not have been produced with-
out pressure from above. 56 Changes in the patterns of surplus
53 SeeChaudhuri 1985:32-33and 182-183abouttheeffectofshon-termshocksontrade.
Quotation from p. 182.
54 The Chronicle of Ps. -Joshuathe Stylite, chapter39, trans. W.Wright,cited in Gamsey
1988:4.
55 Garnsey 1988, chapters 2 and 3.
56 Chayanov 1986 remains fundamental on the productive logic of the peasant household.
Peasants will normally aim to produce a little above their subsistence needs. This
strategy ensures protection in years of less than average harvest results and enables
78
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
extraction are thus likely to constitute a key factor in explaining
developments in economic circulation. Compared to the ecology,
it represents a much more dynamic variable. So, the discussion
has come full circle. We are back with tribute extraction and the
Roman state.
The place oftribute extraction in the economy: Rome and
Mughal Indiacompared
Consonant with the prevailing "minimalist" view of the Empire,
it hasbecome customary to stress the markedly small share ofthe
total economy whichthe Roman state wasableto command. By
themostprobableestimates, it wasintheregionof5-8%ofgross
production. Someareinclinedto goevenlowerandoptfor3-5%
orless.Butthisdoesnotseemrealistic; it hardlyallows anyroom
whatsoever for the state in the economy. s? Historians of Mughal
Indiastruggle withtheopposite problem: thestateseemstotakeup
too much oftheeconomy. Forcenturies the Mughals have served as
the quintessential example ofan oppressive despotism ruthlessly
plundering its miserable subject population downto the very last
penny. It might therefore be useful to confront the experience of
the two empires in order to sharpen our idea of the impact of the
Roman state on the Mediterranean economy.
The Mughal Empire was a large and powerful state based on
the floodplains of northern India. -58 It was founded in the 1520$
byBabur, a central Asian aristocratic warrior adventurer, claiming
descent from the great conqueror Timur Lenk. Sometime ruler of
Samarkand, andlaterKabul,hisluckchangeddramatically forthe
!t.o.'.'lS:..o.[SUPPIiesforbaclyears(Gamsey 1988:54).However, this"normal"surplusis
partofsubsistenceproductionandshouldnotbeconfusedwiththesurplusproducedfor
thepolitical elite^HordenandPurcell 2000,chapter7,inspiteofperceptire comments
on the effects of political pressure on surplus'production, fail to make tha't'cruciai
distinction Wolf1966 chapterI treatsthedifferent"surpluses"producedbypeasant^
^kinsj995-6:,. 46-18for5-7%;Locascio2000C'foIlowmgGoldsm'thI987^^^^^
E'l>'.r.an.!st,imate,of3-5%'butsusPectinglhateventhismaybetoohigh.ThisTsunlikely.'
Duncan-Jones(1994:52^ estimatesthereveniieoftheRomansTateVnigypt"m"the
secondcenturyat 20-25%oftheprovince'sGDP.Thisrevenuealonewouldsufficeto
!:'T8_!tatei"come,CI?setothethreshotdsuggestedbyLoCascio,withEgypt'producing
58 iLdls,?fT'r?onateIy ,largesharec"f', theEmPire'sGDP(seeFurthernoteT76'bd'ow).~
SeeRichards 19938fora survey ofMughiil history.
79
 THE ROMAN EMPIRE AND COMPARATIVE STUDY

better when, in a series of invasions, he succeeded in conquering Extent of Mughalempire. 1530

the remains of the Delhi Sultanate. During the next two centuries Extent of Mughal empire, 1605

the Empire continued to expandanddevelop. Akbarwhoruled Extent of Mughal empire. 1707

Babur's Afghankingdom showing
from 1556toi 605isgenerallyrecognisedastheconsolidatorofthe attempted Mughal expansion
~ ~" T"~ Suri empire
mature Mughalstate.Underhisgrandson,Shahjahan(r. 1628-57), Attempted Mughalexpansion
theEmpireseemstohavereachedits"Antonine"moment embod-
le d Tn Spectacular architectural feats such the Taj Mahal and as

the Red Fort of Delhi. Aurangzeb (r. 1658-1707), his successor,

seemingly continuedthetriumphal marchofMughalarms with
large conquests in the southern parts of India. At the end ofhis
reien. however, strains had built up in the system. It had proven
impossible to consolidate some of the new acquisitions.entered
In the
deca des after the death of Aurangzeb, the Mughal Empire
its "third-century crisis", from which it wasnever to recover. Fre-
quent challenges to the reigning monarch andregional fragmen';
t'ation graduaUy divested the Mughal emperor of his power and
territories many years before the British delivered the coup de Bihar
"^JHARKHANp
graceanddeposedthelastincumbent intheaftermath oftheGreat 'atamau "V

IndianMutinyofT857.
In terms of size and dimensions, the Mughal Empire may
arguably be seen as the most attractive of the potential eariy
modem comparisons sketched out in the previous chapter. The
Ottoman Empire admittedly sharedmuchof.itsterritory withthe
Romans, but thetotal extent, both in terms of geography andpop-
ulation, was significantly smaller. The reach ofMughal territories ARABIAN BA Y Of
wassmaller, too, though with anaction radiusof some 1,200Jan BENGAL
protrudingfromthecapitalDelhi,itwasdistinctlysizeable.-59The
population, however, wasbiggerthantheRoman butstill within a
comparable orderofmagnitude. Bytheturnofthesixteenthcen-^
tury,ltheMughal Empire'is generally estimated tohawcomprised
some loo-no million inhabitants. 60 By then, the Chinese pop-
ulation had already grown to an even bigger size. The Mughals,
therefore, provide'the closest biggerearly modernparallel to the
Romaneconomy (seeFig.2. 1 fora mapoftheMughalEmpire).
5t) Gommans2002: io8. ^ ..... "" ,.^ _"_.."
6° Mo'o's'vFl ^Tchapter estimates 100 million_in 1600. Habib
17 &Raychauclhy^^8^,
for all of India
c:h^pter6'est'i'matet 'l'lo million for the Mugha) Empire and 150 million
in 1600, growing moderately to 200 million in 1800. Fig. 2. 1 The Mughal Empire (Richards, New Cambridge History of India,
1993, p. iv)
8o
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
Despite its reputation as an insatiable Leviathan, the Mughal
Empire also resembled the organisation of the Roman Empire m
several ways. To begin with, it was a conquest state A unitary
system was therefore out of the question. In spite of the impe-
nal facade of homogeneous administration, the system can best
be described as a patchwork, accommodating a large variety of
localandregionalarrangementsofsubmission. 6'Next,bothpoli-
ties belong to the category of patrimonial-bureaucratic empires'
Thismeans thatthe statewasbasedon a large imperial household,
the army and an imperial aristocracy filling important positions
within the system. IntheMughal Empire thetitle usedto describe
the imperial elite was mansabdar, which would be roughly the
equivalent of Roman senators and knights in the imperial ser-
vice and officers in the army. Bureaucratic features, on the other
hand, were not strongly developed. The mechanisms governing
the administrative and military system, for example, depended
more on patronage connections to the emperor than bureaucratic
principles suchaspromotion onmerit, clearly definedlinesofcom-
mandand spheres of authority. 63 Similarly, both empires divided
provincialruleintotwobranches.Thegovernorwouldbeinmili-
tarycommandandresponsibleforkeepingtheprovince loyal and
peaceful. The collection of taxes, on the other hand took place
under the authority ofa separate official, answering directly to the
imperial centre. 64
An exhaustive listing of all similar features is unnecessary.
Clearly, thetwo tributary systems resemble eachother sufficiently
6' On the Mu.ghal Empireandits heterogeneouscomposition, see Richards 1993.^ See
further Aia m"l986; Blake 1991; Hintze 1997; Alam and Subrahmanyaml998;^Habib
'l999^hapter 5. See Richard's 1975 t'or monograph studying the provincial^ settlement
a tax rolls he wanted employed, those of the coward or those of the
i'nGolc'onda.' For the Roman Empire Garnsey and Saller 1987; Jacques and Srt
1990;andLepelley 1998emphasiseorganisational heterogeneity andtheimportai
patrimonialism ratherthanbureaucracy
62 Weber 1972:580-624developedby Blake 1979onthe Mughals. LoCascio2000b:
8'dismisses the notion of a patrimonial-bureaucratic empire. But his own analysis,
pp^.13-79, 97-149and163-176,ofthecomplexitiesoftheRomanimperialpa(rimonit<m
does not convincinglybring it beyondthe model.
63 Blake 1979,AIamandSubrahmanyam2000: 112-130.SeeSaller 1982,chapter3 on
Roman imperial administration. .....,, "__:.
^ The third'book of the Ain-i Akbari, a grand description of the JVfughaj
c. 1600,explainstheprinciplesofimperialprovincialadministratio"-seef"rtherBl'&e,
7979:8'6--90;Richards1986,introductionand1993:5»-59.SeeEck2000onRoman
provincial administration.
82
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
to warrant closer comparison. Of course, there are also impor-
tant differences, the most obvious being that the Mughal Empire
was landlocked whereas the Roman surrounded the Mediterranean
Sea. From the present perspective, however, this rather works to our
advantage. Traditionally the realm of the Mughals has been seen as
the quintessentially land-based, destructive, predatory empire that
kept the economy in a stranglehold. In recent years, this image has
gradually begun to disintegrate and new interpretations, exploring
a more positive connection between economy, trade and empire,
have evolved. 5 At the heart of this historiographical tumabout
is a model which links the collection of imperial taxes with an
increasing monetisation of the economy, much as Keith Hopkins
has suggestedfor Rome. This makes the MughalEmpire an attrac-
tive comparative tool for identifying alternative approaches to the
Roman trading system.
A central concern for historians of the Mughal Empire has
been what one might call "the taming of the revenue intake". Our
sources for Mughal taxation are dominated by official, adminis-
trative records, often of an ideological nature such as the courtier
Abu al-Fazl's survey of the Empire and its institutions, contained
in the late sixteenth-century Ain-i Akbari. Working from these
it seems that the Empire aimed at taking half of the gross agri-
cultural production in taxes. 66 This is incredibly high. It implies
that the Empire came close to confiscating the total agricultural
surplus and sometimes even more. In practice, that must have
been impossible. An anecdote about one of the nawabs of Awadh,
Burhan-ul-Mulk, is revealing about what must have happened.
When trying to increase tax collection, he was asked which set of
man. Only the coward's lists were normally in use and they oper-
ated at a more moderate level of extraction. 7 The Mughal Empire
also seems to have recognised the precarious connection between
official claims and actual revenue. Discounting systems were
65 Habib 1999(1963)is a firststep.Laterresearchhastaken thisfurtherandpartedwith
the notion of destructive despotisms: Bayly 1983, Alam 1986, Richards l993a, Hintze
1997 and Alam and Subrahmanyam 1998.
Habib 1999:230-236.
7 Alam 1986:2I2-2I3;n.28 relates theanecdoteaboutthe tax lists.
83
 THE .ROMAN EMPIRE AND COMPARATIVE STUDY
introduced that allowed the imperial administration to allocate
parts of the revenue, not at the exaggerated (imaginary) official
level but at something which at least for a while resembled the
actual figure more closely. 68Inthe Ain-iAkbari onealsoreadsthe
advice that the level of taxation should be adjusted so as to take
account of the strongly varying agricultural potential and hence
tax-carrying capacity ofdifferent areas. 69A recentestimate, there-
fore. more moderate, but still high, sets imperial tax as one-third
ofstaples such asrice, millet and wheat. More valuable crops were
probably taxedcloser to one-fifth. 70
Several aspects of this can be brought to bear on the Roman
case. The following sections will deal with the imperial state and
the economic surplus and with the process of taxation and elite
formation. Stateandsurplus, first. OurMughalcomparison teaches
us, that 5-7% of gross production is a misleadingly low measure
of the Roman state's impact on society. To be sure, in modem
terms the figure clearly is absolutely low; a welfare state such as
Denmark claims approximately one halfof GDP in tax and still
leaves people more forprivate consumption than everbefore Yet
the Mughal case reminds us that the limits set on the scale of
activities in a pre-industrial economy dominated by agriculture
are much narrower than we are accustomed to thinking of today.
If the dimensions of the Roman state were small, so was the size
of the economy. 71 Agriculture may in general have been a little
more productive on the floodplains of northern India than in the
Mediterranean. Still, pre-colonial Indian conditions provide us
witha roughideaofthespaceweareoperating within.Notmuch
more thanhalfoftotal production, andoften less,wouldhavebeen
available for extraction in the first place (see Fig. 2. 2). 72 Therefore
a more realistic appreciation ofthe scale of state expenditure can
only be gained from seeing it in relation to the total disposable
surplus; It is the size of the disposable surplus which effectively
defines the amount of resources that can be diverted for purposes
68 Alam 1986:24-25. Habib 1999,chapter7 treatsthediscountingmethodsin greater
detail.
<"' Ain-i Akbari, book 3, chapter 7. 70 Richards 1993: 85.
7) Braudel 1981 is fundamental.
72 Moosvi 2000: 33"posits a fifty-fifty scenario forthe Indianagricultural economy arour
1600.
84
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
ES Disposable
surplus
E3 Agricultural
subsistence
Fig. 2. 2 Roughmodel of agrarianeconomy
other than the immediate reproduction of the agricultural sector.
This was the negotiable part of the economy. The rest was spoken
for in advance to cover the subsistence requirements of the food
producers themselves. 73
The implications of this can be brought out with greater clarity
by attempting a simple quantitative model of the Roman economy.
The very attempt at quantification may seem of dubious worth,
given the many uncertainties in our knowledge ofmacro-economic
parameters in the Empire. The results, however, must not be mis-
taken for exact figures. Quantitative speculation cannot generate
the statistics which we do not possess. Nonetheless, as a heuristic
device it can provide us with a rough idea of orders of magni-
tude. Even imprecise figures can be an advantage here. Economies
are finite, quantifiable entities. By giving the Roman economy a
hypothetical, quantitative expression we can control our qualita-
tive analyses with much greater precision than would otherwise be
possible. Quantification, however imperfect, helps us to choose
between competing hypotheses by making their implications
clearer.74 It is also important to recognise that the construction of
this sort of model is not completely arbitrary, in spite of the uncer-
tainties. We do have bits and pieces of the information we need
from the Roman world. This can be combined with what we know
of agricultural economies from other periods and regions to pro-
vide us with a reasonably realistic impression of the Roman situ-
ation. Only industrialisation has created enormous differences in
the productive capacities of complex societies. Agrarianate soci-
eties do show local and regional variations in wealth, but these
73 Haswell 1967: 88; Kula 1976: 62-75; and Garnsey 19985: 22-28 about the centrality
of the surplus.
74 Hopkins 1995-6: 41-46 is a strong defence of the approach. Bang 2002b attempts to
develop the distinction between heuristic device and statistical model.
85
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
fall within a relatively narrow range. And what is more important,
the differences tend to be ironed out when very large areas such
as the Mediterranean, India or China are compared. 75 It follows
thatestimates on a global scale, suchasarerelevant to theRoman
Empire, operate within a fairly manageable margin ofuncertainty,
say 30 or 40%.
'ItakemyleadfromHopkins' recentattempttoquantifythesize
of the Roman economy. 76 The calculation has been summarised
in Table 2. 1. The all important component in an estimate of the
Roman, as of any agrarian, economy is the size of the population.
Duetotherelatively modest differences inproductivity thenumber
ofproducers is goingto bethe singlemost important determinant
of the size of the economy. The population of the Mediterranean
world around 1600 was set by Braudel at 60-70 million. 77 The
number in Roman times is unlikely to have gone far beyond that;
neither Gaul nor northern Italy were as densely populated as later.
Moreover, historical demographers operate with a slowly rising
secular trendbeneath thecyclical pattern ofpre-industrial popula-
tion growth and decline. 78 Taking that into account, Frier's recent
more detailed estimate of the Empire, province by province, can-
not be wide off the mark. Allowing for a slow rise during the first
two centuries AD, he sees the population as reaching a peak of
approximately 60 million in the middle of the second century just
before the arrival of the Antonine plague. 79
From there it is possible to derive an estimate of the mini-
mum agricultural production neededtokeepthispopulation alive.
Modem studies, based on observation in developing countries
and colonies, provide a fairly good ideaof the minimal material
requirements of populations living in agricultural societies. On
average a person needs to consume 250 kg wheat equivalents per
year to subsist at an absolute bare minimum level. Within this
75 Bairoch 1993: loi-l 10andPomeranz2000,chapters I and3.
76 Hopkins 1995-6: 44-48. 77 Braudel 1972: 394-398. ^ _ ^., """.. ":.
7s Scheicfel2000:63-76 suggeststhepossibilityofa maximumpopulation or so miiiiuiL
Bu't"tha~t'does~nonake account ofa rising secular trend through Eumsian pre-industrial
populationhistory.OnthisseeBraudel1981:31-34and88-92andLivi-Bacchi1992:
79 Fner^oooa:812,buildingonBeloch 1886,nowfollowedbyScheidel2007a.Seealso
Frier 20()0b.
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
Table 2. 1 Estimate of disposable surplus, state impact and
interregional trade in the second century AD
i. The population of the Empire in the mid secondcentury AD is set
at 60 million.
2. Minimum subsistence is set at 250 kg wheat equivalent per person
per year.
3. Agriculture was relatively low-yielding. A reasonableestimate puts
average yield at 4 x seed. This means that 25% of total minimum
produce would have to be put aside for next year's crop.
4. Minimum GDP will then be:
60 million x (250 kg subsistence + 83. 33 kg seed)
== 20 million tonnes wheatequivalent
5. A large part of the population will have had to live close to this
level. But cities were normally privileged, some people were also
very rich and not all peasants were equally poor. Production of
manufactured goods and services by the 10-20% not employed in
primary agricultural production also needs to be added. It seems
reasonable with Hopkins to increase total production by half in
order to get to actual GDP: 20 million tonnes x 1. 5 = 30 million
tonnes wheat equivalent.
6. The average price of wheat is set at HS 3 per modius (6. 55 kg). To
judge from prices in Egypt this may be a bit on the high side. But
since this operates against the point I am trying to make it only
serves to strengthen the model.
7. This makes GDP: 30 million tonnes x HS 3/modius == HS 13, 700
million
8. From comparative evidence of other pre-industrial societies we can
operate with a proportion of 10-20% living in cities and 80%
working the land to feed the remaining 20%. The total disposable
surplus can then be defined as GDP minus minimum subsistence
for 80% and seed required to produce subsistence for 60 million.
That is unrealistically generous, but by increasingthe surplus which
actually entered circulation outside the peasanthousehold and its
village network, I work againstthe point I am attempting to make.
9. In numbers this makes our estimate of the total disposable surplus:
[ 3, 700 - (48 million x 250 kg + 60 million x 83. 33 kg) x HS
3/modius = HS 6, 000 million (in round figures) or 40-45 % of
GDP.
io. Following Duncan-Jones,set state expenditure in the middle of the
second century AD at approximately HS 900 million. Direct state
expenditure then constitutes 15 % of the disposablesurplus.
(cont.)
87
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
Table 2.1 (cont.)
ii. To this should be addedthe imperial elite: 600 senators with an
annual income (not to be mistaken for the census requirement) of
around HS i million each, and a similar number of knights, whose
income is estimated at half that of senators. The income of the
imperial elite then can reasonably be assumed to be in the same
range as that of the state: HS 900 million. In other words, the central
aristocracy can account for another 15% of the disposable surplus.
12. io% of GDP, entering interregional (intertown) trade, works out at
23% of disposable surplus.
20% of GDP, entering interregional (intertown) trade, works out at
46% of disposable surplus (add another 10% and we would reach
69% of surplus which wouldbe totally unrealistic).
13. Increasing GDP by one-third would accommodate, in the model,
either raising the price of grain from HS 3 to 4/modius, or setting
population at 80 rather than 60 million, or seeing actual production
taking place at 2 times rather than i . 5 times subsistence. That
would bring GDPup to approx. HS 18,000 million, and reduce the
combined weight of imperial state and aristocracy to 17-23% of
the surplus. Follow any two of these three options and the
corresponding figures would be 13-17% of the surplus. That would
still denote a significantinfluence on the economy, but with GDP at
some HS 24,000 million, the costs of the central government
institutions (HS 900 million) would now only take up a miseriy
3-4% of the total economy.
To conclude, any understanding of economic integration in the Empire
must needs take the imperial tribute extraction as its point of departure.
It is able to accountfor a very substantial part of all interregional flows
of resources, be they in the form of trade or of redistribution.
figure food accounts for around four-fifths, the rest is provision
for a little clothing, firewood, housing etc. It should, perhaps, be
stressed,thatthesearereal minimums, presupposinga low level of
activity, rather thanthe somewhatgenerous prescriptive standards
of calorific intake recommended by United Nations agencies. 80
The notion of wheat equivalents may require a little further
explanation. Ideally, of course, we would like to have statistics for
all the different products that people consumed apart from wheat,
in order to determine their relative significance in the economy.
80 dark andHaswell 1970,chapter I andpp. 57-62.
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
In the absence of that, wheat equivalents are a useful fallback.
Before industrialisation food grains play a dominant role in the
economy of the vast majority. As a consequence, the value of the
rest of production tends to stand in a rough relationship to the
price of grain. In times of famine, for instance, the value of non-
essentials drops sharply because people need to draw heavily on
their resources to procure highly priced cereals or their substitutes.
Therefore it makes good sense initially to express an agrarian
economy in terms of wheat equivalents. 81
Thus, by combining population size with subsistence require-
ments minimum production in the Empire can be set to
60 million x 250 kg = 15 million tonnes wheat equivalent. The
imperialeconomy,however,couldnothavefunctionedatthislevel.
Twomore factors shouldbe addedto the calculation. First, it takes
seed to grow grain. Now the seed : yield ratio is far from an uncom-
plicated expression of productivity. High yields may for instance
be obtained by a low sowing rate, making the return per hectare
more modest. For our present purposes we need not go into the
finer details of these problems. What we need is a rough measure
and 4 times seed has been adopted, acknowledging that behind
this figure there were enormous local differences. We do not pos-
sessmuch goodinformationonyields in theGreco-Romanworld,
but the scraps we do find in the various agricultural treatises do
not strongly contradict this choice. Moreover, scientific analysis
of ancient grains found by archaeologists confirms that they will
have been fairly low-yielding. 82 Finally, on comparative grounds,
wecanalsoarguethatyieldmusthavebeenhigherthanthree-fold
because otherwise the peasants would have found it difficult to
support the elaborate urban-centred social system. On the other
hand, they would probably also have been below five-fold or we
would have to accept a much higherlevel ofurbanisation than most
arewillingto concede.83So4 times seedcanbeusedwithreason-
able confidence as a broad measure. This adds an extra 5 million
" Gamsey 1988:3-7 citing thechronicle ofJoshuatheStylite on a famine in lateRoman
Edessa.Forcomparativeconfirmationoftheroughrelationshipbetweenpricesinfood
andnon-e.ssentials,seeChatterjee 1996:54.
In addition to Hopkins, seeGamsey 2000: 706-709 ongrainproductivity.
De Vries 1984: 242-243 works with a standard peasant household of 5 members and
8 hectaresofland.Withyieldsof3 : i thehouseholdwouldproducelessthanits own
89
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
tonnes wheatequivalent, andsubsistenceproduction in the Empire
comes out at 20 million tonnes wheat equivalent (Table 2. 1, steps
3-4).
While many peasants,village artisans andpoor townsfolkprob-
ably existed at or close to subsistence level, considerablenumbers
had a standard of living above this. 84 Cities, for instance, were
privileged places where large resources were concentrated. It is
not uncommon to hypothesise city populations as consuming agri-
cultural products at twice the minimum level. 5 Some of this will
have been used as raw materials by urban manufacturers. This
reminds us that we also need to include the production of higher-
value goods and services performed by the 10-20% of population
living in the cities as well as manufacturing in the countryside. All
in all, our minimum estimate of the economy probably needs to be
increased by half in order to arrive at the real GDP. This brings
our estimate of the Roman GDPto 30 million tonnes wheat equiv-
alent. Finally, this number must be converted into monetary terms
to enable comparison with state expenditure. Our information on
wheat prices in the Empire is fragmented, but suffices to give a
broad impression. 86 In most parts of Egypt the price of wheat
seems to have been around HS 2 per modius, at least until the late
subsistence. At 4 : I the household would be able to support 1. 4 persons employed
outside agriculture. With yields rising to 5 : I the number of persons would go up to 34
or about 40% outside agriculture. Erdkamp 2005: 34-54 argues, perhaps a little one-
sidedly, for the possibility of higher yields in the Mediterranean, especially Italy, with
a potentially even higher surplus to sustain aon-agricultural activities. But, as he also
emphasises, his model farm is a Utopia, designedto draw attention to the significanceof
peasant agriculture for creating a low-yield regime. The model "high-yield" farm would
only be possible on the best of lands, and could not reproduce its own labour force. In
other words, it depended on the availability of a peasant population in the surrounding
countryside. Average productivity would thus be much lower than the hypothetical
maximum.
84 See Gamsey 1998, chapter 4 on the near subsistence existence of many groups in the
Greco-Roman world.
85 Habib 1999:84-85. See further De Vries 1984: 242-243, whoobserves that it is rela-
tively more demanding on resources to maintain people in cities than in the countryside.
ForHellenisticBabylonanimpressivecontinuousseriesofpriceshavebeenpreservedin
the astronomical records. Aperghis 2001 makes use of this material to model the macro
conditions of the Seleucid economy in Mesopotamia. Unfortunately, he has forgotten
to include production above subsistence, which weakens his results considerably. For
example, the state is seen as confiscating60% ot'GDP,even though that means that the
peasants are left with less than their subsistence requirements. He also opts for a level
of urbanisation at 50% of population. That seems extremely optimistic. See Aperghis
2004 for a broad study of the Seleucid economy.
90
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
second century. Similar prices are attested for Pisidian Antioch
(HS 2. 25), a small town inAsiaMinor, andRomanAfrica. Onthe
otherhandprices in Rome were probably more like HS 8-10per
modius. 87 Hopkins haschosen HS 3 ashis average. If anything,
that is probably on the high side, since most .grainwasconsumed
outside thelarge urban concentrations wherefoodwasmore expen-
sive. Using a price of HS 3 per modius brings our estimate of GDP
in the Roman Empire in the mid second century roughly to HS
13, 700 million (Table 2. 1, steps 5-7).
Admittedly therearegreatuncertaintiesinvolvedinthecalcula-
tion. But as a rough measure I think it will do. Stated in per capita
income it works out at HS 229. 88 Some years ago, an economist
estimated per capita income in the Empire in the first century AD
as HS 38o. 89That is problematic. This figure wasreached without
using any significant analytical checks to assure that the estimate
wouldberealistic.As anaverageHS380,infact, seemsunrealisti-
cally high when we compare it to what we know ofRoman wages.
Adjusting for differences in wage levels between city and coun-
tryside and between men and women, it implies that the average
male working income was higher than the pay of a legionary in
the first century andonly a little less in the secondcentury. Thisis
implausible. Thelegionarieswerea powerful andprivilegedgroup
in Roman society. We would expect them to do considerably bet-
ter than the average adult male labourer. Consequently, per capita
income should be comfortably below HS 380. Indeed, if the infor-
mation preserved about the wages of Roman labourers is used as
an alternative basis from which to calculate a possible GDP, we
arrive at a per capita income a little below our estimate. 90 It looks
as if the dimensions of the model are roughly correct.
87 DuncaiWones 1982: 144-147 andRickman i98oa,chapter6 collect much oftheevi-
dence. The normal price range was HS 2-4. Egyptian'wheat prices are collected in
Dunt-a"-Jones1990:151;Drexhage1991;andRathbone1997.
I." whlatecluivalentsthefigureis500kgperpersonornet'productionofseed417kg.
In 1960,India,still heavilydominatedbya very basicagriculturalregime,hadana
pej;caPltaagrlculturalproductionof382kgwheatequivalent(darkandHaswelI1970:
78). This lends credence to our estimate.
89 Goldsmith 1987:35.
I, lmmII,chob^ged to Richard saller t'or presenting this argument to me in correspon-
d.en';e'. t,'?i"g Dunc:an-Jones '982: 54SaIIersetsurbanmale labourat HS3 aday'and
!lur. .at. HS .I;;'', He sets population 60 million with half actually working (250 days
at
peryear),20%livingincitiesandfemaleincomeathalftherateofmale,a practicethat
91

THE ROMAN EMPIRE AND COMPARATIVE STUDY
We can now use our GDP as the basis for an attempt to quantify
the disposable surplus, the really crucial element in the Roman
economy. In this exercise, it is important to reach a generous
estimate, more likely to be an overestimate than the opposite, since
this will work against the point I am trying to make. Therefore I
havetriedto includesomeanalytical"buffers"whichwillbeableto
absorb most of the effects on our conclusion of an increase in GDP
of, say, 10or20%. Again, taking thebetter-known situation around
1600 as our starting point, urbanisation seems to have reached a
new high of 16-17% of the Mediterranean population. 91 Figures
for the Roman Empire are unlikely to have been significantly
higher, but we allow a generous 20%. This leaves 80% of the
populationtoproducetheagricultural surplus.Todothat,itneeded
its minimum subsistence requirements and seed for producing the
minimum GDP. All production in Roman society above those
two elements, I take as the disposable surplus. This is, in fact,
very generous. Peasants also had to feed their draught animals
and would probably in many areas have been able to reserve a
little more than the bare minimum for the consumption of the
household.92 Leavingthis aside,the disposablesurpluscomesout
atroughly HS 6,000million. Basically this meansthat only about
40-45% of the Roman economy would have been available for
taxation, market circulation and so on (Table 2. 1, steps 8-9).
Roman government expenditure in the second century AD can
be estimated at approximately HS 900 million. 93 Combining this
estimate with our GDP enables us to see that the net expenditure
of the imperial system (mainly the emperor's household, "bread
is attestedfromantiquity.Finallyheassumesthatthesexdistributionofthepopulation
was fifty-fifty. This works out at HS 10,000 million. To this he adds20%for elite
incomeand7-8%forstateincome.ThefinalresultisapproximatelyHS 13,000million
orHS217percapita.IfweinsertGoldsmith'sestimateofHS380intotheequationwe
see, since only half of the population worked, that average working income would be
twice thebasefigure, thatis760. Ifmale income isdouble thatoffemale, averagemale
working incomewould thenbeHS i ,013.Forcomparison theannualpayofa legionary
was HS 900 in the first century andHS 1,200 in the secondcentury (Duncan-Jones
1982: lo). Temin 2006 goes even lower in his estimate, alsobasedonprices, but with
less analytical sophistication.
1)1 DeVries 1984:69-77.
92 dark and Haswell 1970, chapter 4 argues that of production up to 350 kg wheat
equivalent perperson, most will usually remain within thepeasanthousehold.
93 Duncan-Jones 1994: 45; Hopkins I995-6; Walters 1999: 202-234.
92
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
and circuses", and the army) probably took up at least 15% of
the disposable suq^lus, as opposed to only 6% of gross produc-
tion. This is quite substantial, especially when we remember, as
pointed out by Hopkins, that the state often spent its tax income
at some remove from where it had been collected (Table 2. 1, step
lo). 94 If, for example, we hypothesise that interregional transfer of
resources amounted to something like 10% ofGDP, c. 65% could
arguablybe accountedfor by the flows ofmoney, goodsandredis-
tribution in kind caused by net tribute extraction. If the figure for
intermediate and long-distance exchange of resources is increased
to 20% ofGDP, the proportion comes down to 33%. But at 20% of
GDP intercity transfers will amount to more than two-fifths of the
disposable surplus (Table 2. 1, step 12). Since a large part of the
surplus was undoubtedlyconsumedin its region of origin, interre-
gionalexchangecannoteasilyhavebeenmuchhigher.I will return
with some more detailed reflections to support this view later in
the chapter. For now it will suffice to observe that even in what
seems a "worst case scenario", net government expenditure is still
far from a marginal economic force. Thus as a preliminary conclu-
sion it seems clear that the imposition of imperial tribute cannot
easily be bypassed in our understanding of the development in the
intermediate and long-distance transfer of resources. At the very
least, it must have significantly modified the pre-existing pattern in
the Mediterranean world. The Roman state may after all not have
been quite so marginal an economic presence as our GDP figure
initially indicated. In that way it is moving closer to our Mughal
example.
Tribute extraction, elite building and
economic integration
Our comparison with the Mughal Empire, though, still confronts
us with a glaring discrepancy between the size of the two states:
the Roman state probably took up less than 10% of GDP whereas
the Mughal state controlled perhaps up to one-third. That is curi-
ous and should make us wonder. In terms of organisation, we
94 Hopkins 19786; 1980; and 1995-1996.
93

THE ROMAN EMPIRE AND COMPARATIVE STUDY
remember, the Empires looked much alike. Could they really have
experiencedsuchdifferingsuccessrates?Ifwecomparetheimpe-
rial armies, by far the largest items of state expenditure in both
realms, we would not expect such a stark contrast. The army of
the Great Mughal seems to have comprised around 200, 000 cav-
airyand40,000infantry,whereastheRoman,consistingmainlyof
infantry, counted approximately 400, 000soldiers attheendofthe
second century AD. 95 Even allowing for the greater expenses con-
nected with the equipment of armed horsemen, the Mughal army
is still surprisingly small, especially when we addthat the sizeof
theMughal population wasarounddoublethatofthe Roman. 96
Apparently, thegreatershareofthesurplus ina muchlargerecon-
omy didnot easily translate into military striking power ofvastly
different dimensions. Something is missing from the equation. We
forget that we are comparing netexpenditure ofthe Romans with
gross income ofthe Mughals. Taxes never come in cheaply. Being
without a large, disciplined, but costly modern-style bureaucracy
to effectuate their commands, the Mughals instead had to create
loyal andpowerful aristocratic groups atvarious levels ofimperial
society to perform the functions of government. In return, these
groups weresupported bytheassignment ofvery substantial shares
of the tax revenue.
Thus the Empire came to reshape the structure of north Indian
elite society. At the baseofthe imperial formation alliances were
formed with village headmen and local bigwigs, the so-called
zamindars. ^ Their position in the countryside was strengthened
astheywereleftinchargeofcollection andpayment oftheimperial
taxes. This enhanced their control and authority over the peasantry
and enabled them to increase their incomes, often substantially,
from their share in the imperial revenues and from tax rebates
or exemptions of substantial areas of land under their direct con-
trol. not to mention from the increased opportunities to impose
illicit or unofficial extra "taxes"; a Mughal letter formula for the
appointment ofthe headman for a district of villages includes an
95 Hassall 2000: 320-321 and Duncan-Jones 1994: 34-
96 Moosvi1987,chapter17;HabibandRaychaudhuri1982,chapter6.
97 Bayly 1983;'Alam1986,chapter3;Richardsl993a:86-91;All 1997:84-87;Hintze
1997,chapter6; andHabib 1999,chapters4-5.
94
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
admonitionto the appointeenot to makeanychargesunsanctioned
by local custom.98 Clearly, even officially,to saynothingof actual
practice, the state had to recognise considerable room for "cre-
ative" interpretation by its various representatives. The price the
local elite groups had to pay for their increased powers was greater
dependency on the state for their position. Though still in a posi-
tion of substantial autonomy in the day-to-day running of things,
openly disloyal, rebellious or overly recalcitrant wmindars risked
facing the Mughal army and replacement by a more manageable
substitute, for instance a local rival."
The increased flow of surplus production out of the Indian coun-
tryside, brought about by this alliance of local and state power,
went into financing the central imperial institutions: the large
household of the great Mughal himself and the imperial, military
nobility, the so-called mansabdars, who controlled and recruited
the largest part of the cavalry army. Of those two the lion's share
was taken by the mansabdars. It has been estimated that they were
allocated about 80% of the enormous revenues of the Empire,
whose collection they had to organise through their agents and in
collaboration with the local elite groups. Most of this, probably
around 60% or so, was intended as salary for the soldiers they
were obliged to maintain. The rest mainly consisted of a personal
payment for their services rendered towards the great Mughal and
was intended to finance their aristocratic and luxurious lifestyle.
Thus the Mughal Empire can be likened to a giant revenue pump
at the end of which was created an incredibly rich imperial elite
with large tbllowings organised around their grand households
dominatingthe main cities of the Empire.
9 Richards 1986:41. Impositionoftaxesis neverneutralin relationto theexistingsocial
structure. They are bound to invest their administrators with increased powers which
will be used to their private benefit. Hinton 1966, chapters 2-4, an anthropological
micro-study of the Chinese village of Fanshen, provides one of the most revealing
descriptions of the repertoire of coercive tools available to a gentry in control of state
taxation for bringing the peasantry into economic dependence and for the eventual
confiscation of its lands.
99 AIam 1986, chapter 3, especially 92-93; Richards l993a: 86-91; Ali 1997: 84-87;
Habib 1999, chapter 5, especially 208-222.
'"" Richards l993a, chapter3, especially75-78.SeefurtherQaisar 1967andMoosvi 1987.
See Blake 1991 for grand households.
95
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
As sketched here, much in this process of intertwined state and
elite formation seems rather familiar from what we know about
the development of the Roman Empire. Perhaps, it is time for a
more inclusive understanding oftheRoman imperial state orrather
a less unitary one? One peculiarity of the Mughal arrangement,
though, would normally be taken to contradict this. The majority
of themansabdarswere not allowedto form more enduringlinks
with the countryside which could be transformed into permanent
landholdings. Every so often their revenue assignments would be
reshuffled and they were allocated new areas from which to col-
lect their taxes. 10' This contrasts with the Roman world, where
the resources of the imperial elite, consisting of senators and
knights, were invested in enormous conglomerations of private
landholdings. Ever since the famous French traveller and dealer in
diamonds, Jean-Baptiste Tavernier (1605-89), reported back to a
horrified European public on how the so-called despotic practices
oftheMughalsdeniedthenobility therighttoprivateproperty, this
has been taken as the defining difference between the powerful,
independent more feudal forms of aristocracy found in the West
and the weaker, state-dependent Oriental nobilities. '02 Therefore
the tribute extractionof theRomanstate is generally describedas
being structurally opposed to the independent rent-seeking activi-
ties of the landed elites. They are perceived astwo separate forms
ofpowercompeting forthelimited surplus production ofthepeas-
antry. 103
There is no denying that the Empire knew a structural conflict
between, for example, senators and emperor, about the distribution
of the surplus, but so did Mughal India. An important concern of
the imperial establishment was to prevent the mansabdars from
maintaining fewersoldiers thanthey wererequired to whilepock-
eting the saved expenses for their own benefit. 104 The ultimate
101 Richards 1993: 66-68. Seefurther Alam 1986, chapter I; Ali 1997, chapter 3; Hintze
1997,chapters4 and8; andHabib 1999,chapter7.
.
°2 Tavemier 1889: 260 (book II, chapter l).
.
°3 Thus Hopkins2000: 254-257 and Rathbone2000.
104 Ali 1997: 53-59 and Habib 1999 discuss this question but tend to underestimate
theextentot thephenomenon;evenif themansabclarmaintainedthefull numberof
soldiers he might still pocket a profit if he could hire the soldiers for less than the
awardedamount,e.g. Richards 1993:64-65.
96
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
control claimed by the Mughal emperor over the landholdings of
the imperial elite was a tool which enabled him to punish and
reward individual members of the imperial aristocracy. It could
not, however, be used to reserve all revenue for the emperor him-
self; he had to allow the aristocracy effective control of much of
the imperial revenue. The regime of the great Mughal depended
on his ability to attract the service of great and powerful men.
Such men expected to be granted areas from which to collect rev-
enue in return for their loyalty. The Mughal emperor had to assign
to his aristocracy most of his notional rights to agricultural rev-
enue in a form of prebends called jagirs, a term which is perhaps
best described as denoting a kind of fief with the restriction that
the monarch retains a greater degree of ultimate control than was
normally the case in feudal Europe. A failure by the monarch to
secure access for his aristocracy to collect the wealth of extensive
agricultural areas would plunge Mughal rule into crisis. '05
The contrast between the two systems, one based on private, the
other on prebendal, aristocratic land tenure, can, in other words,
easily be made too stark and risks concealing a more fundamen-
tal similarity. In many ways, observing the Mughal system, where
tribute dominated rent-taking, provides a tmer picture of the work-
ings of the Roman system. 106 Certainly, the accumulation of vast
property portfolios in the hands of the Roman imperial elite was
intimately connected with its access to the resources generated by
the acquisition of empire in the Mediterranean. It is in no way clear
that the Roman imperial elite was more independent and powerful
than its Mughal equivalent. The mansabdars may not generally
have been allowed the luxury of vast private landholdings, but
then they constituted the army and administered most of the impe-
rial revenues. Had they in that position been conceded even more,
one may seriously doubt that there would have been an empire
rather than a string of petty principalities. On the other hand, the
Roman senate, in spite of its landed wealth, usually found itself
105 Richards 1975; Pearson 1976; All 1997 (1966): 92-94.
Cf. Haldon 1993: 63-87 who argues against Wickham 1985 that aristocratic rents
should not be seen as structurally distinct from tribute extraction. Wickham 2005: 57-
60 now accepts this, but with an important modification: only as long as a centralised
state maintains an ability to tax the peasantry.
97
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
the weaker party when it confronted the emperor and/or the army.
It was that experience which gave the historical work of Tacitus
both its impassioned verve and disillusioned detachment. In times
of dynastic turmoil, it was not in the senate that the new emperor
was made; strong support in the army was overwhelmingly the
decisive factor. When the courtier and senator Nerva was put on
the throne in Rome after the murder of Domitian in AD 96, his
feeble hold on power was only secured through the adoption of
one of the foremost generals of the Empire, Trajan, as heir to the
throne. 10?
Nothing illustrates better the relative weakness of the senatorial
order as opposed to the imperial system than its manifest failure
effectively to make membership of the senate hereditary. As shown
by Hopkins and Burton, even in the case of the more successful
senators, who rose to the highest offices and obtained the consul-
ship, only a third managed to have male descendants within the next
three generations reaching a similar standing. Some families would
have died out in the male line. But a significant number would just
have withdrawn from politics in the capital altogether. '08 Partici-
pating in politics at the imperial court was a high-risk activity. The
expenditure alone required to maintain a senatorial household and
lifestyle could be crippling to any fortune, however big. '09 Most
senators would have found it necessary to add to their income the
substantial profits that could be made from holding high offices
in the imperial administration, such as army commands or gover-
norships of provinces. 110 Indeed, senators expected to be allowed
a substantial share of the resources generated by the imperial sys-
tem. The senatorial historian Dio Cassius, for instance, makes
Maecenas advise Augustus to take the leading magnates across
the Roman world as "partners in empire". There is a distinct busi-
nesslike flavour to the choice of vocabulary. Later in the Roman
History, Dio criticises his own emperor, Caracalla, for giving his
freedmen too great a share of the imperial proceeds while the
107 Syme 1958, chapters I and 2 for NervaandTrajan. SeeRoda 1998: 213-221 in general
about the impotence of the senate. Cf. Tac. Hist. I, 4.
108 Hopkins 1983(1: 134-146. I09 Hopkins 1983(1: 149-175.
''" Suet. Vesp. 4. 3 is quite telling about this. Here it is a cause for comment that Vespasian
refrainedfrom enrichinghimselfduringhis governorshipof Africa, so endangeringhis
credit. Most aristocrats knew better, cf. Brunt 1990, chapter 4.
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
senators received nothing. '" It is the same pattern of thought we
find behind Pliny's description of Trajan, donating and selling
off parts of the imperial estates to members of the Roman elite;
the princeps is complimented for allowing his friends (amid) to
benefit, instead of treating the Empire as identical with his own
household. "2
Ofcourse, that was precisely whattended to happen. Being so
closely woven into the fabric of the imperial system, the fortunes of
the elite were highly vulnerable or permeable. The great narrative
histories of the early empire of Tacitus and Dio Cassius provide
ample illustration of greedy emperors killing senators and other
aristocrats in order to confiscate their property. Yet, it would be a
mistake to see this only as a question of periods of terror under
a few mad and cruel rulers. The phenomenon had much deeper
roots thanthis, andfoundits truest expressionin a setofpractices
common to "good" and "bad" emperors alike. During the early
Empire the custom developed among members of the elite and
various sections of the imperial apparatus to include the princeps
as co-heir in their wills. Furthermore, the emperor appropriated
the right to properties left without an heir, without a valid will, and
those belongingto people condemnedon serious chargessuch as
murder or offences againstthe state. "3 Here is how Pliny, in his
Panegyric to Trajan, elegantly labours to turn black into white,
pressure into heartfelt gratitude:
You are not named as heir because someone else has given offence, but on your
own merits, set down by your personal friends and passed over by strangers. The
onlydifferenceinfactbetweenyourformerprivatelifeandyourpresentsupreme
position lies in the greater number of those who love you, as your own affections
aremorewidelyspread.Onlycontinueonthiscourse,Caesar,andexperiencewill
show whether the reputation and purse alike of a prince are not better enriched
when his subjects make him their heir at their death from choice rather than from
compulsion. "4
Dio Cassius 52, 19, i-3 uses the expression "xoivcovoi CTOI Tf)s apxr)s". The term is
also used about proper business partnerships, e.g. Liddell and Scott, Greek-English
Lexicon s. v. I, 2. For Caracalla, Dio Cassius 78, l8, 4. See further Cassius Dio 52, 5,
1-2;52, I2;52,28 abouttheexpectationsofsenatorsofsharingin theproceedsofthe
empire. Millar 1964: ro2-i 18 and chapter5 is still valuable.
112 Plin. Pan. 50. "3 Millar 1977: 153-174.
114 Plin. Pan. 43, citation from paragraphs 1-4, with Millar 1977: 156-157and Saller
1982:7I-73.
99
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
Onegets a very clear sense oftheharsher reality behind this glossy
rhetorical image when Pliny goes on to describe it as a favour if
theemperoracceptsbeingpassedoverina willbysomeonewhom
he hadbenefited in some material way or other.
Anyabsolute distinction betweentheprivate fortunes oftheelite
andtheimperialstatethereforeseemsmistaken.Theestatesotsen-
ators and knights equally belonged to the world of high politics.
The connection comes out in full clarity during times of dynastic
collapse. Theensuing struggles forthethrone andchanges inpower
relationships werenormally accompanied bya redistribution ofthe
wealth of the elite on a considerable scale. Aristocrats linked with
the losing factions risked facing destruction, confiscation oftheir
property andits subsequent redistribution among themembers of
the victorious coalition, many of whom would have been stimu-
latedbythedesireforspoilsorbytheunsettledstateoftheirprivate
affairs"."5 It istellingthattherateofhereditarysuccessionamong
the highest-standing and hence best-established consulars seems
to have fallen from around 50% to only 30% in times of political
instability. [ I6Basicallythen,theenormousfortunesoftheimperial
elite depended upon having access to the power resources gener-
atedby the empire andultimately controlled by the emperor. "7
This reflected the fact that they were also very much the creation
ofthe empire. The period ofrepublican expansion hadwitnessed
thebuild-upofproperty andwealthinthehandsoftheRomanelite
ona steadilyincreasingscale financedbyprofitsdrawnfromth^
exploitation andconquest oftheconstantly expandingEmpire."
The growth of aristocratic property portfolios continued under
the Principate. It wasfuelled by a stronger integration ofprovm-
ciallandholdingsintothetributary systemoftheRomanstate.The
"5 Tac. Hist. [I, 7. Thecloselink betweenviolentdynasticstrugglesandredistribution
ot aristocraticwealthisa mainthemeot'Tacitus' Histories,e.g.1, 2;I, 20, I, 88;md
II, 84.Of'course oneshouldbecautious oftakingTacitus' laments too seriously. The
sug~gestion,however,thataristocratswithfinancialproblemshadeverythm8tojam
from'the turmoil surely isquiterevealing, notbecausewe should acceptTacitus' sneer
thatallthe"troublemakers"werebankrupts,butbecauseoftheunderlyingexpectation
thatsuchpeople wouldbeabletoprofitsubstantially fromjoining a rebellion.
"6 Hopkins i983a: 138-139, including table 3. 6.
"7 Sal'ler 1982,chapter2 andMillar 1977:275-355showtheintegrationot theelitein a
network of imperial patronage.
'I8 Hopkins 19780: 39-56. Seefurther Shatzman 1975: i i-[76.
100
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
process went in two directions, from the centre to the provinces
andfrom the provincesto thecentre. Alreadyestablishedmembers
of the imperial elite would to a larger extent than hitherto acquire
estates scattered around the provinces. As provincial administra-
tors, they profitedfrom the dispensingofjustice andthe collection
of taxes by various more or less sinister methods such as accep-
tance of gifts, "bribery", pocketing of parts of the imperial rev-
enues and collection of extra dues. 'I9 Some of this would have been
transformedinto provincial real estate. Throughprosopographical
studies it is possible to follow howsome senatorialfamilies, based
in Italy, gradually built up large estates in parts of the provinces
where they had governed. 120 An important aspect of this process
was the lending of money at high rates of interest by wealthy
Roman aristocrats to provincial communities, lacking money to
pay their taxes. '21 Seneca is said to have lent something like HS
40 million, at a very high rate of interest, to needy communities
in the newly conquered province of Britain, only to have the debt
called in shortly after. The harshness with which he proceeded to
collect the debts was supposed to have been an important con-
triburing factor to Queen Boudica's rebellion. 122 Undoubtedly, an
exacerbating factor was that many had seen their lands confiscated
in the process of settling the debt. In the late fourth century the
Syrian orator Libanius complained that aristocrats in state offices
exploited their positions to lend money to town councillors and
then took over their estates whenpayment failed. 123
A central element in all of this was the easier access enjoyed by
people of high standing to the coercive machinery of the Empire.
Plutarch's description (Moralia 8l4D) of the leading imperial offices in the provinces
as "TToAuTaAavTous"(generatingmuchmoney) is revealing. Onprovincial administra-
lion: Saller 1982, chapter 5; Woolf [998, chapter 2; Eck 2000: 281-288. Specifically
on taxation: Neesen 1980; Garnsey 1988, chapter 15; Merola 2001.
Roda 1998: 214-215 with bibliography.
Howgego 1994. Dig. 18. 5. 9 discusses the selling-up of land burdened with tax arrears.
122
123
Dio Cassius 62, 2, I.
Lib. 48, 3 and 14. Though late, Libanius is no less relevant for the earlier period. The
practice was ingrained in Roman provincial administration. Apart from Seneca, it is
vividly attested in the correspondence of Cicero. The letters, Alt. 5. 21; 6. 1; 6. 2 and
6. 3 (with Rauh 1986), discuss Brutus' usurious loans to provincial communities. See
further Plutarch's essay, "That we Ought Not to Borrow", in the Moralia, paragraphs
828E-829E on the general importance of money lending for the redistribution of land
in provincial communities.
101
 T
THE ROMAN EMPIRE AND COMPARATIVE STUDY
Senators, knights and other persons of substance would often, for
instance, have been able to use their connections within the impe-
rial system to secure the support of government troops to settle
conflicts to their own advantage, not least in connectionwith the
collection ofdebts. 124The importanceofthis shouldnotbeunder-
estimated. Roman society wasnot strongly regulated andpeaceful.
Violencewasendemic.Peoplewho suddenlyfoundthemselves in
weaker positions, such as orphans and widows, frequently expe-
rienced difficulties in holding on to their material possessions. 125
Claims needed to be defended and asserted in order to count. 126
This was also true of the legal system. Hearing a case, as Aulus
Gelliusinformsus,wasnotjust a matterofweighingtheevidence.
The general standing and social clout of the litigants were equally
important matters to take into consideration. 127 A trial was to a
very large extent a question of measuring out the relative power
and influence of the contestants. '28 Hence the established prac-
tice ofusing one's powerful connections to influence the outcome.
Again it does not require much imagination to see how a Roman
senator or knight with a stake in the financial exploitation of the
provinces would benefit from this. To judge from the collections
of letters preserved from antiquity, it was an aspect of the mutual
exchange of favours inside the personal networks of the imperial
elite to write to provincial governors with whom one had links,
asking them to look kindly on the cases of one's friends, proteges
and connections. Ofcourse, it is difficult to know how much influ-
ence such recommendations exercised on the governor/judge, and
it will in any case have varied from case to case, according to the
circumstances. But the very fact that these letters were written,
shows that people expected them to have an effect. 129 Thus, it
seems reasonable to suggest that the various kinds of support that
the imperial elite could mobilise from the state system must have
124 Brutusattemptedtoobtainmilitarysupportforthecollectionofprovincialdebtsfrom
CiceroasgovernorofCilicia,Cic. Att. 6. 1.5-6. Seefurthertheanecdotalevidencein
Apuleius' Metamorphoses (VII, 6-7) and MacMullen 1990: 191-193.
[25 Krause 1994, chapter 10 and pp. 232-235 and Krause 1995, chapter 13, especially
pp. 195-200.
126 Millar 1981: 66-71 and MacMullen 1990, chapter 18.
.
27 Gellius 14.2 withGamsey 1970:210-211. I28 Saller1982:56-60.
129 Saller 1982: 152-154.
102
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
played a crucial role in enabling it to acquire and maintain estates
in different regions of the Empire. '3"
Another way forward was to team up with a local aristocrat
and town councillor who enjoyed sufficient influence locally to
promote and guard the economic interests of rich outsiders. Liba-
nius, for example, complains about how the leading members of
the town councils as middlemen for people in government posi-
tionshandledruinous loansthatdeprivedtheirless prosperouscol-
leagues oftheir landholdings. '3' Undoubtedlythese local bigwigs
profited substantially from their involvement in those transactions
with people in higher positions. As such, this only represents a par-
ticular instance of a more general phenomenon. Plutarch urged the
leading town councillors to cultivate the patronage of members
of the imperial provincial administration because "the Romans
themselves are most eager to promote the political interests of
their friends" and would serve them "as a firm bulwark". 132 The
reason for this was, of course, that the Empire depended on the
payment of taxes. But being without a large bureaucratic corps, it
needed to create a group of local allies who were strong enough
to control the area, ensure its continuous peaceful subjugation and
handle the actual collection of the imperial tribute. In his auto-
biography Josephus, a local Jewish grandee, provides us with
a close-up of how this was done in practice. When the Jewish
rebellion had been almost quelled, Titus, the Roman commander,
offered to spare Josephus' friends and relations from punishment
andcaptivity, thuspreservinghisnetworkofsocialcontacts. After-
wards Josephus was presented with lands in Judaea, first by Titus
and then later by his father, the ruling emperor, Vespasian. Later
still, the emperor Domitian added to the gift by granting tax immu-
nity to his properties in Judaea. Throughout this period of imperial
favouritism Josephus also enjoyed the support and protection of
the emperors against political rivals who tried to bring him down
by accusinghim of treason againstthe Roman cause. 33
]3° Saller 1982, chapter 4 for the importance of access to the resources generated by the
Empire to the imperial elite.
131 Libanius 48, 37-41. I32 Plutarch Moralia 8140.
133 Josephus Vit. 422-430. The autobiography is itself a defence written to counter the
accusations of a rival, Ju.stus of Tiberias (Vit. 336-367, 410). Brunt 1990, chapter 13
on Josephus and Empire.
103

THE ROMAN EMPIRE AND COMPARATIVE STUDY
Josephus, however, was also exceptional. The vast majority of
local aristocrats would never have experienced anything resemb-
lingtheclosedirectpatronageoftherulingdynastythatheenjoyed.
Yet on a scale less grand, the privileged cocktail of protection,
land and tax benefits was offered to those in powerful positions,
who where willing to co-operate, around the provinces of the
Empire. 134 In organisational terms, the Empire was only lc)osely
integrated. As has been said so many times before, the world of
the Romans was, rather, a patchwork of conquered city states.
Thesewereretainedat the time ofconquestasthebasicandinter-
nally largely self-governing administrative units. In areas where
city state organisation had been lacking, the Roman conquerors
introducedit in orderto satisfythe needsofthe imperial adminis-
tration. Among the functions of city government was the respon-
sibility for the payment of the imperial taxes. Each city tem-
tory was assessed at a set amount of tribute, and the leading and
richest citizens were effectively left in charge of distributing the
burden internally within the landowning segment of the citizen
body. 135Obviously, this gavethem considerable power androom
for manoeuvre. To begin with, they were normally able to ensure
that their own property would be underestimated, one way or
another. 136 Moreover, this did not just enable them to avoid pay-
ing their full share of the taxes. Usually, it also meant increasing
the burden on those less well off and hence less able to shoulder
it. 137 Thus an avenue of access was opened up for the extension
of rural credit. We have already seen how this was used as an
investment opportunity by rich outsiders, but it was also open
'34 Some general references must suffice: Jones 1940; Gamsey 1988,chapter 15; Brunt
1990,chapter 12; JacquesandScheid 1990,chapters5-7.
'35 /GV(l), 1432-1433 (with Hopkins 1995-6: 68, n. 28) showhow thecity of Messene
allocated its'tax of"100,000 denariiamong its citizens. See further Bay 1972: 32-47;
Jones 1974, chapter 8; Neesen 1980; Brunt 1990, chapter 15.
I3f> P. Beatty Panop.2, lines68-71 and 145-152is a goodexample.Theleadingmag-
natesm'Panopolis hadcolluded with a village official to leave some cultivated land
undeclared, thereby effectively carving outtax-free landsfortheirownbenefit.
'37 dc. U Verr.2, n8 isclearaboutthismechanism.SeeVeyne1990:200(andnote292)
whoalso addsthe testimony of CodexJustinianus,Xl, 58 (57), I: a law of AD 313
whichobservesthat"thetabulariiofthecities,incollusionwiththepowerful,shiftthe
burden of taxes onto the lesser people".
104
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
to the prosperous insiders themselves. Plutarch admonished the
leading citizens not to (ab)use the support of the Roman author-
ities to satisfy their own greed and drive away the less power-
ful. 13 Libanius adds greater detail to the picture. Much to his
dismay, the local landed magnates did not just act as middlemen
for people in higher positions. They also took advantage of their
power actively to increase their own wealth by imposing too heavy
burdens on their lesser colleagues in the city councils and after-
wards forcing expensive loans upon them that made them lose their
lands. '39
True, both Plutarch and Libanius criticise the system. In that
respect, they are also witnesses to ideological attempts to limit the
most brutal effects of its workings. The Roman jurists, too, gave
rulings that prohibited the uneven distribution of tax burdens and
government requisitions. 140 Such ideological misgivings and legal
admonitions, however, should not be interpreted as fundamental
attempts to combat and change the basic principles of the exploita-
tive machinery. Plutarch, after all, recommended municipal rule
by the leading and richest members of each community, supported
by Roman government representatives. Rather, these phenomena
should be seen as (feeble) mechanisms to contain those activities
within bounds in order to ensure the longevity of the system. The
sheep had to be sheared, perhaps even fleeced, but not gutted. Too
harsh a regime might undermine the tax-paying capacity of the
cities by causing social unrest, banditry or just impoverishment
and abandonment of cultivated land. It was a question of main-
taining a balance. But that balance remained heavily tilted towards
the members of the most prominent families around the cities of
the Empire because they were the most important guarantors of
social stability. 4
n!i Plut. Mora;;a8l5a.
!39 Libanius 48, 37-41. The practice of overburdening weaker members of the councils
is confirmed by rulings in the Digesta, e. g. Dig. L.4. 3. I5 (Plutarch Moralia 827-832
warns more generally against the corrosive effect of credit on the economic fortunes
of councillors). Control of taxation and loan-giving was a powerful tool of wealth-
acquisition in the hands of pre-industrial elites. Hinton 1966: 50 is illuminating.
140 D(g. L.4. 3. l5(Ulpian).
4 Cic. Q. fr. I, 1, 25 is clear about the need to avoid internal stmggles in the cities,
banditry etc. to maintain the tax capacity of the province.
i05
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
One gets a very clear idea of this from the institution of courts
to hear complaints of provincials about the activities of Roman
governors. In Cicero's speeches against the most notorious culprit
in provincial maladministration, Gaius Verres, a recurrent strategy
of discrediting the accused is to emphasise how he alienated and
encroached on the wealth of the leading members of the Sicilian
provincial communities. Verres' crime as governor was not just
his deeds of self-enrichment, but more the damage done to those
upon whom Roman rule was based. In seeking to profit from his
office Verres had not respected the established social hierarchy.
Therefore, according to Cicero, he could only muster the support
ofsmall fry andpersonsofdubiousreputationandpowerwhohad
gained from his attacks on the homines nobilissimi primique^
Whether a governor would be prosecuted and punished for his
(mal)administration seems, in other words, to have been decided
less by what he did than whom he did it to. It was essential that
he secured support from a section of the provincial upper classes
with sufficient strength to be able to block attempts to have him
indicted after his time of office had expired. 143
As in the Mughal Empire then, the imposition of Roman hege-
many considerably modified the pre-existing pattern of power
around the numerous cities of the empire. The administration of
imperial tribute extraction enhanced the strength of the leading
local landowners. As a result, these elite groups were able grad-
ually to accumulate wealth and property on a larger scale than
before. 144 A clear sign of this is that "big men" with very large
financial capacities indeed begin to surface in our sources. The
province of Lycia, on the south coast of Asia Minor has revealed
some of the best evidence for this process through its very strongly
developed epigraphic habit. A whole dossier of inscriptions, recov-
ered from a funerary monument, describe the brilliant career of
142 dc. II Verr. 2, ll. Further II, 2, 14, 22, 35-49 dismissing provincial supporters of
Verres as small fry anddesperate people hopingto benefitby turning the social order
upside down.
143 CIL XIII. -(162, with Saller 1982:132.
144 See Brunt 1990, chapter 12 and Garnsey 1988: 246-251 for this development. For
studies of individual provinces, seee. g. Garnsey 1978 (Africa), Millet 1990(Britain),
Jones1940;Quass1993;Merola2001;Dmitriev2005(GreekEast),Woolf1998(Gaul)
and 0rsted 1986 (Noricum and Illyricum).
io6
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
a certain Opramoas of Rhodiapolis.145 Normally attention has
focused on the enormous gifts that he bestowed on the Lycian
communities during a period of a quarter century (AD 125-52).
They seem to add up to something like 1,000,000 denariior about
four times the minimum legal requirement of a senatorial for-
tune. 146 This is munificence on a truly staggering scale. In Paul
Veyne s interpretation, Opramoas is the ideal embodiment of the
spirit ofeuergetism, a virtuoso taking to an extreme the habit of the
aristocracy of making public benefactions by performing financial
suicide in a kind of ancientpotlatch. '47
I remain sceptical about that interpretation. Opramoas was not
alone in the province and neighbouring regions in demonstrating
such spending capacity. '48 Ultimately, of course, in the absence
of his private accounts we will never know whether he did indeed
ruin his patrimony. But what we do know is that gift-giving on this
level, sustained for many years, requires a substantial build-up of
wealth and the generation of very large incomes. Altruism is not
nearly enough to maintain such activity. Here we find Opramoas
and his family, exactly as we might expect, playing a central role
in the political life of the province and cultivating the patronage of
Roman magistrates andemperors. According to the inscriptions the
family had a prominent record of office holding within the Lycian
koinon, or league, perhaps stretching back before Roman rule. 149
But the first member for whom we have more detailed informa-
tion is Opramoas' father Apollonios. He played a dominant role
in his hometown Rhodiapolis, holding all its major offices, and
attained a position of some prominence in the provincial coun-
cil, though he does not seem to have reached the highest office
of lykiarch. 150 However, he had certainly managed to concentrate
enough wealth and influence during his career to see two sons,
including Opramoas, off to even greater glory and influence. We
meet Opramoas first in AD 125, when his father performed the
office of archiphylax both in his own name and in the name of
145 TAM 905. Kokkinia 2000 presents a new improved edition of the text, with detailed
"real kommentar". Heberdey 1897 reconstructed the chronology of the information
contained in the inscriptions.
14(' SeeCoulton 1987: [72 fora recentdiscussion. 147 Veyne 1990: 149-150.
148 Coulton 1987. I49 TAM 905, [VC, lines 13-14.
150 TAM 905, I1B, IIIE-Fand VIIIC.
Toy
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
his two sons. This was an office which required great financial
resources. In the honorary decreesOpramoasis describedas hav-
ing paid "the taxes owed to the imperial coffers by the province
with his own money and only afterwards began the collection with
greatphilanthropy". '51 Onewouldliketoknowtheeconomic real-
ity behind this phrase. But in the light of the foregoing discussion
it seems farfrom unlikely that Opramoas or, to bemore precise, his
father, would have been able to profit from handling theprovincial
taxes. If nothing else, we find the family securing the patronage of
the governor lulius Frugi, who praised them in official letters. 152
At any rate, the patrimony of the family was certainly not
ruined orendangered bythis. Onthecontrary, during thefollowing
twenty-five years Opramoas (and his brother) expanded the power
of the family and rose to the highest positions in the province,
including secretary and lykiarch, the leader ofthe provincial coun-
cil and high priest ofthe imperial cult. All of this brought him into
close contact with successive Roman governors. Their many letters
of praise addressed to the provincial council allow us a glimpse
of how Opramoas continued the work ofhis father and expanded
his network of imperial supporters. 153 Several of the governors
even assisted in bringing Opramoas to the attention oftheemperor
AntoninusPius,whohonouredhim in letters on morethan a dozen
occasions.154This does not look like a man eroding his economic
151 TAM 905, IIE, lines 10-14. From the proceeding information andtext 14 and XIIIC,
line 4, it is clear that his first positions were held through his father.
'52 7-AM905.IF.ContraQuass 1993: 177-178.whoonlyseestheeuergetisticsideofthe
handling of taxes.
'53 Opramoasenjoyed the patronageof effectively every governor in Lyei a duringthe
period.Weknowthe nameofthegovernorin twenty-twoyearsofthe twenty-eight
years covered by the inscriptions, and all of those more than ten governors found it
worth their while to support Opramoas - one or two of them, though, only after some
initialresistance.Oncompletion ofthe lykiarchyOpramoaswasvoted extraordinary
honours.Askedtosupport'this,thegovernorcrypticallyrepliesthatthekoinonoughtto
follow established practice. This isprobably notsupport, asKokkiniatakesit tobe, but
ratheranattempttocurbtheambitionsofOpramoas.Thefollowinggovernorcontinues
theresistanceto theseextraordinaryhonoursbut is outmanoeuvredby Opramoaswho
securesthebackingoftheemperor(TAMgos,VIIA-VIIIA,Kokkinia2000:225-228).
Rather than a "potlatch" aristocrat, Opramoas better fits Tacitus' category (Ann. xv,
xx) of leading provincial magnates whose great wealth allowed them to dominate the
province and were of crucial importance to the governor.
154 TAM 905, texts 37-51 areletters from the emperor honouring Opramoas. Several had
beensupportedbythepatronageofgovernors.Heberdey1897:70-71hasa convenient
list.
io8
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
base by excessive expenditures. Instead, the inscriptions convey
the impression of a family steadily increasing its resources and
sphere of influence. A further indication of this is that as his career
progressed he seems to have acquired considerable interests in
the province outside his hometown Rhodiapolis. Some of the later
inscriptions show him as having acquired citizenship in several
other Lycian communities. He is also described as pursuing poli-
ties actively in all the cities of Lycia- a fact which is documented
by the very wide distribution of his benefactions across a broad
selection of the cities in the province. 155 They could, of course,
be interpreted as an instance of pure generosity, but it seems more
reasonable to take them as a sign of Opramoas' need to cultivate
his interests on a province-wide scale. He had now grown beyond
his hometown.
In this connection, it is certainly not without interest that the
family continuedits expansivemove in the next generation. One of
the last inscriptions proudly records that a nieceof Opramoashad
enteredthe imperial elite by marryingthe senator, ClaudiusAgrip-
pinus. '5 The couple is also known from a fascinating inscription
found in the Lyciancity Oenoanda,listing a genealogy of another
leading Lycian family. '57 From that we can see that Agrippinus
eventually reached the office of consul. In the generations follow-
ing this union we find senators, consuls and even a governor for
the province ofLyciaandPamphyliafrom this family. Agrippinus
himself was from an old family of lykiarchs. One of his forefa-
thers was lykiarch as early as around the middle of the first century.
Two generations later we find Agrippinus' grandmother marrying
a Roman aristocrat, who held the governorship of Spain and also
commanded a legion. The daughter of this marriage married a
lykiarch and in the next generation we then find the son, Claudius,
and the daughter, Claudia Helena, re-entering the imperial
aristocracy.
As a group the inscriptions show in detail how the interac-
tion of local elites with the tributary empire gradually led to the
155 TAM905, VIIIB-Ementionssixcitieswithoutbeingexhaustive.Texts53 and59show
the wide dispersal of Opramoas' benefactions.
156 TAMgos-XVIIID,lines 12-13.
57 IGRIII,500. SeeHeberdeyandKalinka 1896for a graphicrepresentation.
log

THE ROMAN EMPIRE AND COMPARATIVE STUDY
formation of stronger and richer elite groups. At the pinnacle of
this elite, are to be found some families which had risen to a level
of wealth that made them big even in a provincial context and
whose interests would often have exceeded the bounds of their
local community. This made them attractive partners for already
established members ofthe imperial elite, both interms ofmarriage
and as resourceful allies who could be promoted into the highest
orders of imperial society as they invariably needed replenish-
ment. '58 Detailed and painstaking prosopographical research into
the composition of the Roman senate has revealed how during
the first two centuries AD under the auspices of the emperors, it
changed from a body of Italian aristocrats to a gathering of the
"best"men" from across the Empire. '59 This is, of course, a sign
of the relative weakness of the senatorial aristocracy. But it also
reveals the strength ofthe imperial system. It managed, asit were,
to (re)fashion and (re)produce its own elite. 160What is more, this
was done on a steadily increasing scale as larger and larger accu-
mulations of provincial land were merged into the existing stock
of wealth in the hands of the imperial elite. Thus, to talk about
the weakness of the imperial aristocracy is paradoxical, for it grew
richer and came to command larger social resources through its
dependence ontheimperial system. LiketheMughal Empire, the
Roman Empire appears to have worked like a powerful revenue
pump.
Tributary integration: themarket assurplus transformer
Thus the comparison with Mughal India adds a new perspective
to our understanding of patrimonial bureaucratic empires. The
's8 Saller1982,chapter5 onthepatternsoftheprovincialpatronageoftheunperialelite.
'w The exoression is commonplace in the contemporary literature: Aristeides Or.^
2^4^'7'l"and'DioCassiusr52, 15 with Oliver 1953; Millar 1964: . 02-118^md
Aapte'r 5"ChastagnoFl992 a"dKoda 1998: 172-213, chapter n survey±%reseal;ch
on the changing composition of the senate. Stein 1927, chapters 3-4 and ?>rlerw1"^
White l973a, e.g. 402-417 are still valuable Hammond 1957 is tu"clame"taL,
furtherAli-5ldyi977;Hiilfmann1979;Epigrafiae ordinesenatono
and
1982. Particularly
on knights, see Pflaum 1950; Demougin 1988; Demougin Deyi jver 1999. ^
.
60 A'c'l'eaTexpressionofthiswasthecreationofa secondlegoftheimPerial^elite^as
knightswe'readmittedtopowerfulpositionsintheEmpirenexttothesenate;seeSal
1982,especiallychapter3;Hopkins l983a:[76-184.
110
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
customary emphasis on a conflict of interest between tributary
empire and various rent-taking aristocratic strata is only one part
of the imperial experience. The empire is not just engaging in a
socially destructive war against its aristocracy where it will either,
as a capstone in John Hall's phrase, curtail and break the landed
elites or disintegrate into numerous small cells. Rather, the central
power depended for its existence on partly dissolving the oppo-
sition and successfully integrating the more local forms of power
into the imperial fabric. Some years ago the late Burton Stein, a
historian of pre-colonial India, put forward the concept of seg-
mentary state to denote this kind of political system. By this he
meant that the imperial state could not be reduced to the cen-
tral institutions of the emperor's household and the army, it also
encompassed "a large sub-stratum of local lordships, magnates
or... honoratiores". '6' The great advantage ofStein's formulation
is that it recognises that state (tribute) and local forms of power
(feudal rent) are not absolutely separate. As aristocracies come to
participate in the functions of the state system, through office hold-
ing and the collection of taxes, the two forms of authority begin
to blend. Invariably they become intertwined and, depending on
the success and durability of the imperial formation, mutually
supportive.
Tribute extraction, in other words, was never just a question
of depleting the resources of local communities. In return for
handling taxation, aristocrats were, as we have seen, allowed to
profit substantially at various strategic points in the process. This
needs to be added to the bill if we are to reach a proper assess-
ment of the economic impact of the Roman state. At least for
the central aristocratic groups, senate and knights in the impe-
rial administration, it is possible to get an idea of how this works
out in our quantitative model of the imperial economy (Table 2. 1,
step 11). Based on Pliny's letters, Duncan-Jones has estimated the
annual income (not to be mistaken for the census requirement)
of a middling senator at approximately HS i million in the early
Stein 1985(1, especially 387-388 and 410-411, citation from 411. See further Stein
1998: 18-30, stressing the duality of conflict and dependence.
Ill
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
second century AD. 162 Some senatorial households would have
experienced inuch greater incomes. On the other hand, some also
had problems even in meeting the economic requirements of a
political career. 163 Hence it seems reasonable to adopt the figure
based on Pliny's experience as an average annual income for the
6oo or so senators, even though we might actually be underestimat-
ingthewealthofthesenatethen. '64Thecontoursoftheequestrian
tranche of the central imperial elite are more hazy. But to judge
from the growing number ofequestrian procuratorships (127 are
attested in the reign of Marcus Aurelius) it would have required a
group more or less equal in size to the senate to fill the spaces. 5
The wealthiest and most powerful knights clearly were richer than
many members of the senate. But this would not generally have
been the rule, rather the opposite would have been the case if the
lower property qualification of this group is anything to go by. A
fair guestimate would put the income of the equestrian members
of the central imperial aristocracy at half that of the senate. In
numbers this gives 600 x HS I million + 50% = HS 900 million
or an extra 15% of the disposable surplus to be addedto the 15%
which was estimated above for central government expenditure
(Table 2. 1, step 11).
This makes the Roman Empire look a lot more like the Mughal
Empire, especially if we take into consideration that the Mughals'
notionalintakeofone-thirdofproductioninreality wouldprobably
have been closer to one-fourth or one-fifth. Then when we
subtract the substantialpart that went into the pockets of the local
aristocracies, the wmindars, the two empires come to look very
much alike. The revenue of the central imperial system seems to
Duncan-Jones 1982, chapter I. See also Hopkins 2000, n. 5 for some comments on the
material informing us of the scale of aristocratic wealth.
163 Millar 1977: 297-300. 4 Chastagnol 1992,chapter [0.
15 Whittaker 1993,chapter)2: 61-62 adoptsa similarfigurefor thesizeoftheequestrian
section of the imperial elite. Pflaum 1950 and 1960-1 are the classic studies of the
equestrian imperial procurators.
166 The Mughal revenue intake is a thorny issue. But independently of how we assess this
problem two things operate in favour of lowering towards one-fourth or one-flfth of
GDP. The figure of one-third is of agricultural production. To be strictly comparable
to our Roman figure, it should be adjusted to take account of urban production and
services which only contributed modestly to Mughal taxes. On top of this, many cash
crops were only taxed at a rate of one-fifth. Taken together these two factors go a long
way to bring the Mughal tax figure within "our" range.
112
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
have been within a range of 10-20% of GDP in both cases. This
does not look impressive. But when we realise that this would
have constituted close to one-third of the disposable surplus in
the Roman case (and it cannot have been wildly different in the
Mughal case) it is clear that it could not easily have been much
larger (Fig. 2. 3 is a graphicrepresentationof the model). '67
As already mentioned, the revenues of the imperial system
(emperor, armyandcentralaristocracy)werespentunevenlyacross
the empires and often at a considerable distance from their collec-
tion point. IntheRomancaseit wasto a very largeextentconsumed
by emperor andelite in the capital and in the borderregions where
the army was stationed. 168 This means that much of the revenue
intakehadto enterregionalandinterregionalcirculationin orderto
reach its point ofexpenditure. In Table 2. 1, step 12, 1 have tried to
indicate the consequences of different proportions of GDP going
outside its nearest marketing network, that is, entering intercity
resource flows. At 10% ofGDP, about 20-25% of the disposable
surplus would be spoken for in this way. At 20% of GDPit would
be 45-50%. Clearly this must be pushing against the maximum
limit of whatcould have happened. Even though the consumer-city
model has been hotly debated, few people would doubt that a very
large section of the surplus was spent locally by the municipal
aristocracies around the numerous ordinary cities of the empire.
Data from both eighteenth- and nineteenth-century India and late
imperial China, too, suggests that we should envisage a level of
intercity exchange of between 10 and 20% of GDP in the Roman
Empire. 9
167 Huang 1974: 159-i61 on MingChinashowsthenarrowlimits forimperialtaxation.
168 Whittaker 1993, chapter 12: 61-62 for a balanced account of imperial aristocratic
consumption.
lb9 Yang 1998: 223-224 lists a number of estimates from eighteenth- and nineteenth-
century India. They indicate a rise in the marketable surplus from 20 to 30% of
production. Ofthis a significantportion is bound to have been sold withinthe region
ofproduction.Oneestimatesawabout 10%ofagriculturalproductionenterwidernet-
worksofcirculation. Expressedasa proportion ofGDP,whenmanufacturedproducts
and other luxuries are added into the calculation, this suggests a level of interregional
trade of about 10-20% of'GDP. Perkins 1969: 115-120 estimates total value of inter-
regional trade in late imperial Chinaat 15-20%of gross farm output. Expressedas a
proportion of GDP (farm output plus production and services in the cities) the figure
will be a little lower, perhaps 10-15%.
ii3
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
53 State net
expenditure
0 Imperial
elite
E3 Remaining
surplus
E3 Agricultural
subsistence
Fig. 2. 3 Roman GDP, surplus and tributary empire
For our understanding of the economy of the Empire this sce-
nario has important implications. To begin with, it is difficult to
identity market exchange as the dominant organising force ofeco-
nomic integration in the empire. At our imagined maximum level
of interregional resource transfers, expenditure related to the cen-
tral imperial system would be able to account for up to two-thirds.
In practice, some of this would have gone into property trans-
fers from tribute payers to the tribute-receiving groups in society.
This is the implication of the processes of increased concentra-
tion of agricultural land in aristocratic hands described above.
But even in that case, a large part of the politically extracted sur-
plus would still have had to be exchanged between the different
regions of the empire. '7" No wonder that archaeologists again and
again emphasise the different scale of activity between Greek and
Roman history. 171 Imperial tribute extraction must have dramat-
ically changed and expanded or at least considerably modified
the pre-existing pattern of economic circulation in the Mediter-
ranean world. If we want to speak of economic unity in the Roman
Empire we would do well to start by looking for tribute extraction
and the shared processes of elite formation and concentration of
landed property rather than a conglomeration of interconnected
capitalist markets. Consumption related to the profits of empire
comes out as such a powerful stimulus that there cannot be much
room for a development in interregional trade driven by regions
170 See Jongman 2006: 248-50 for an attempt to model the scale of property transfers
spawned by the imposition of an imperial tribute.
171 Mattingly and Salmon 2001: 3-14.
ii4
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
specialising in single products and exchanging them with one
another. '72
This may be stated with confidence. Even if we should have
underestimated the value of the economic surplus and interre-
gional trade in our model by, say, a third, the conclusion is not
much affected; imperial expenditurewould still be able to account
for up to 50% of our maximum level of intercity transfers. This
would, for instance, be the case if the average price ofgrain rather
than the HS 3 adopted in our calculations was closer to HS 4 per
modius. But that would haveto beconsidered a highprice in many
regions of the empire, acceptable perhaps in times of threatening
famine, but not characteristic of the prevailing conditions. 173 A
similareffectwouldbe producedby increasingtheestimateofthe
Roman population from 60 to 80 million. However, that popula-
tion figure represents the plausible maximum. It is as high as the
more optimistic of modern considered estimates go. Ofcourse, the
compounded effect of adopting both modifications to the model
would be more serious, though far from damaging. Such a pro-
cedure would reduce the combined weight of imperial state and
aristocracy in the surplus to approximately 15%. That would still
be a very noticeable, but admittedly weaker, force in the imperial
economy (see Table 2. 1, step 13 for a summary).
But why would we choose to compound our most optimistic
estimates? It is certainly not necessary in order to accommodate
.
a more optimistic" view of exchange in the empire. A model of
an agrarian economy which provides for some 10-20% of GDP
to be transferred between localities does not present a "minimal-
ist" scenario of Roman exchange activities, quite the reverse. It
bears emphasis that the different parameters adopted in the model
cannot simply beincreased atwill. One ofthe advantages ofquan-
titative speculation is that it enables us to observe how the different
parameters are interrelated. Increasingthe value of GDP, without
changing our estimate of state expenditure, has the questionable
Cf.Perkins1969:I2&-I23onthepatternofinterregionaltradeintheChineseEmpire;
the politically determinedconsumption ofemperorandhiselite dominatedoverinter-
regional specialisation.
Cf. the edictof the governorAntistius Rusticusproclaiming i denariuspermoclius
ot grain (twice the normal level) as the maximum permitted price during a famine in
PisidianAntioch(AE 1925, 126= Sherk 1988, no. 107).
TI5
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
implication of reducing the weight of government institutions to
almost nothing in the economy. In the present case of a compound
increase in grain price and population size, the costs of impe-
rial court and army are reduced to 3-4% of GDP. That would be
difficult to justify. '74 The model already operates with a "small-
scale" Roman state. There is little reason to make it even smaller.
Only a generation ago, leadinghistorians suchasJonesand Brunt
thought of Roman taxation as oppressive. '75 The Roman emperor
was the single biggest landowner in the empire; and the fiscus
drew substantial rents from these extensive properties. A recent
estimate of taxation in the province of Egypt sees the Roman
authorities claiming perhaps some 20% of provincial economic
output. Taken together, these two observations alone would suf-
fice to bring Roman government income within the range of 3-
4% of GDP, thus precluding the need for any further taxation
around the rest of the empire. To this list must also be added
income from precious metal mines and stone quarries. It is diffi-
cult to fit the Roman state within the narrow limits of 3-4% of
GDP. '76
174 For some general reflections, Hopkins 1995-6 (more accessible in Scheidel and von
Reden 2002: 201-203).
175 Brunt 1990,chapter 15;Jones 1974,chapter8.
176 Exact figures cannot be hoped for, of course. But 'crown' lands, as in most other
pre-industrial monarchies, are unlikely not to have constituted several per cent of
agriculturalproperty.Egypt,however,allowsatleastsomehypotheticalreflection;it is
the province whichprovides us with the most abundant evidence of levels of taxation
and"production. Working from this Duncan-Jones (1994: 52-55) has estimated the
revenue claim of the Roman state in Egypt in the second century to be in the area of
HS 250million. Veryroughly thiscanbeestimated as20-25% oftheprovince's GDP.
Egyptwasanextremely productiveprovincewithgrainyieldsbeyondcomparisonin
the' Roman world. It also had a rich luxury trade passing through it. If we estimate
Egypt'sGDPasatleast10%oftheimperialtotalwecannotpossiblybegoingtoohigh
rathertheopposite. This meansthatthe taxclaim on Egyptalone will haveequalled
at least 2% of imperial GDP (one-fifth of 10% of imperial GDP). As already noted,
Duncan-Jones (1994: 45-46) estimates government expenditure in the second century
to be around HS 900 million. This is about 3. 6 times the value of the Egyptian tax
figure which equalled a minimum of 2% of GDP. HS 900 million, and hence state
expenditure, will then have beenaround 7% of GDP.This is, at least suggestive. The
figuresareveryuncertainandthemethodofcalculationistoosimplistic.ThusHopkins
1995-6:44 findsDuncan-Jones' estimate ofEgyptian revenue implausibly high partly
because Hopkins overestimates the proportion of this revenue in the state budget
partlybecauseheforgetsthata substantialpartoftherevenuewaslikelyto"disappear'
betweencollection point and state consumption. However, there is no doubtthatthe
Egyptian revenues were of extreme importance for the Roman state.
116
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
In other words, though we do not and never will know the exact
sizeofthe Romaneconomy,our speculativemodellinghasto oper-
ate within certain constraints. Increasingone parameter may well
require us to reduce another in order to keep the dimensions real-
istic. Perhapsthe most problematic parameterin the quantification
of the Roman economy is the choice of multiplier from minimum
subsistence to GDP, set by Hopkins to 1. 5. One cannot claim that it
could not possibly be any bigger. But it could not be much bigger.
The limits canbeexploredby setting up a simpleequationdescrib-
ing the living standards which are possible for different population
segments underdifferent productivity/incomeregimes (how many
times subsistence). Table 2. 2 explains the mathematics and sum-
marises the results of using different values for the agricultural
population, for economies in the range of 2-3 times minimum
subsistence. Let us examinethe implications of the most prosper-
ous scenario sketched in the table, 3 times minimum subsistence.
If 80% of the population was engaged in agriculture then it would
be able to have a living standard at 2. 5 times minimum subsis-
tence while the 20% non-agricultural/urban could live at 5 times
subsistence. Make the agricultural population 60% and the corre-
spending figures would fall to 2 and 4 times subsistence. With the
agricultural population constituting only 40%, it could still live at
1. 88 times minimum subsistence and the non-agricultural popula-
tion at 3. 76 times. This does not seem a realistic possibility for the
Empiretakenasa whole,whichmixedrichandpoorregions.Three
times subsistence would imply that the entire Roman world was
approaching the level of productivity which was achieved (excep-
tionally) in England by the middle of the eighteenth century, one
of the most prosperous pre-industrialeconomies on record, where
only four out of ten had to be employed in agriculture. Stateddif-
ferently, it wouldrequiretheRomaneconomyasa wholeto sustain
a level of production whichwas probably not even achieved in its
agriculturally most prosperous region, the fertile, Nile-inundated
province of Egypt. '77 Both conditions are implausible.
177 ThestudybyD.ThompsonofthePtolemaicvillageKerkeosirisrevealedanagricultural
regime with 7 aroura available per family. With yields at TO times seed, and I artaba of
wheat sown per aroura, then if all land for convenience's sake is imagined to have been
planted with wheat, total net production would be 63 artabai. A family of five would
ll7
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
T
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE

Table 2. 2 Living standardsfor different groups ofpopulation would suggest that the plausible range for the Roman economy
under different productivity regimes would have to be placed somewhere between i and 2 times mini-
mum subsistence. [7SThis locates the hypothetical maximum at the
Let a be the proportion of the population engaged in agriculture, then lowratherthanthehighendofourtableofalternative,moreoptim-
the non-agriculturally employed population is (i-a); and let x represent istic scenarios for Roman living standards. Pressin. g the Roman
the basic standard of living for the agricultural population and allow the
economy to a maximum of 2 times subsistence would as much
non-agricultural population to live at twice this level. S equals
subsistence and y is the number of times subsistence at which per as double my estimated surplus (though in practice probably less
capita production takes place. since it was calculated generously). The imperial state and aris-
2x(i-a)+xa=y * S tocracy wouldthen accountfor a minimum of 15%ofthe surplus.
x (2-a) = y * S As already stated, that would be far from a negligible influence
x = y/(2-a) * S on economic circulation; it would still represent a very significant
change from conditions before the impositionofEmpire. But this
Livingstandards(x) expressedastimes subsistencefor the agricultural/ low result would haveto be consideredan absolute minimum and,
(urban) population at different production levels
as explained, an unlikely estimate of the share of the imperial
(y) Production at: system in the overall economy. Within the plausible limits for the
(a) Population 3 times 2. 5 times 2 times Roman economy, therefore, it would be difficult not to ascribe to
in Agriculture subsistence subsistence subsistence imperial surplus extraction, the role of a key stimulus of interre-
gional economic flows.
8o% 2. 5" (5) 2. 08 (4. i 6) 1. 67(3. 34)
6o% 2. 14(4. 28) 1. 79(3. 58) 1.43(2. 86) This stimulus, however,didnotplay itselfoutonly or primarily
40% 1. 88(3. 76) 1. 56(3. 12) 1. 25(2. 5) withina state-redistributivesystemofgoodsinkind,suchasmight
be envisaged along the line of thought of Polanyi. The imperial
At 3 times subsistence a scenario of only 40% of the population state, for one thing, as already argued, did not havethe organisa-
employed in agriculture begins to look plausible. Standards of living for tional capacity. Markets were a necessary intermediary. Taking a
the agricultural population can be maintained at a reasonably generous last glance at ourMughalcomparison, wecan saythatin the kind
level while the non-agricultural population does very well. Both groups oftributary system describedhere,the role ofmarketexchangeand
would in fact be marginally better off than a population with 80% in
agriculture with per capita income at 2 times minimum subsistence. interregional trade is not first and foremost that of co-ordinating
The latter scenario, however, would still be a reflection of fairly wealthy and organising the economy in an interregional division of labour.
conditions far away from an agricultural minimum state, and most Ratherits task is the subordinateone of actingas transformerand
likely represents a hypothetical maximum for the Roman economy.
11S Macltiison 2001, 46 (figures for the "developed"world). In general chapter I and
appendix B. I suspect that Maddison underestimates production both in the ancient
When we look at the Roman Empire in its entirety, comparisons world and in early modern China. The European figures, however, are more reliable.
of averages are more appropriate on a continental scale, rather Herehe presentsa developmentrisingfrom a real minimum percapitaof$400around
AD 1000 to approximately twice that level in [700. In general, as North 2005: 91
than with smaller, particularly prosperous regions. Estimates of observes from these data, the difference between rich and poor countries fluctuates
average per capita incomes for pre-industrial Europe as a whole between I and 2 times. Jongman 2006: 241 is right in guessing 2-3 times subsistence
asthe upper limit forproductivity. Butthe largerthe geographicalareaunderconside-
ration, the closer the maximum average will approach 2 times. The richest agricul-
require some 32 artabai for a very minimal own consumption, a family of four some tural regions weregenerally fairly circumscribedandappearagainsta more extensive
24 artabai. Total production would therefore be between 2 and 2.5 times minimum background of considerably less productive regions. Saller 2002: 259-261 places the
subsistence. The Roman world could not possibly have exceededthat. Romaneconomy in a bandbetween 1.2 and 1.5 timessubsistence.
ii8 ii9
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
conveyer of the extracted surplus. 179 The majority of tax payers
were grain-producingpeasants, but the requirements of the impe-
rial state were much more diverse. Neither the Roman nor the
Mughal government anddifferent layers of imperial elites, would
have been able to put to good use all the food grains which they
potentially commanded. "The good husbandman", as even that
arch-proponent of traditional agricultural virtues, Cato, admon-
ished, "should be a seller, not a buyer". Therefore government and
aristocratic elites required the services of markets and traders to
mobilise parts of the suq?lus by converting agricultural produce
into money which could be stored and used later in other contexts
to buy different products. In the Mughal Empire, a significant part
of the taxes came to be remitted in coin rather than kind. Something
similar was suggested by Hopkins for the Roman Empire. But in
bothEmpires,this processofcommercialisingthe agricultural sur-
plus is likely to havebeen uneven. Some, even many, taxes would
still have been paid in kind; not all peasants would have been heav-
ily involved in market transactions. Often, the peasant would have
handed over part of his production to bigger landowners or tax
collectors, who would then have undertaken to cart the produce to
market, near or distant according to circumstances. Alternatively,
the landownermight pass on the produce to merchants at the farm
gate or even unharvestecl on the field. 18' The imperial government
would also frequently have availed itself of markets to convert
excessproduce into monetary wealth. Instalments of the Egyptian
grain tax found their way onto the private market in Puteoli, the
harbour where the imperial grain fleet arrived from Alexandria
before the construction of better port facilities at the mouth of the
Tiber. Even surplus medicinalherbsfrom the Cretanestates ofthe
emperors were sold on the market through commercial agents. '82
179 Bayly 1983: 63-68. This is the frequently overlooked implication ofHopkins' taxes-
and-trade model.
Is° Hopkins 1995-6(cited in Scheidelandvon Reden2002: 216-17).
'8' Cato Agr. ,quotation from 11.7 ("patrem familias vendacem, nonemacem esseoportet");
advance sales: CXLVI-I1:sale of wine injars: CXLVIII.
182 Galen XIV, 9 (on the medicinal herbs), TPSulp. 45, 46, 51, 52, 79 (on Alexandrian
wheat used for commercial speculation in Puteoli). Lo Cascio 2006 for the imperial
state as a large seller in the market.
120
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
Merchants and traders, to borrow an expression commonly used
by Mughal historians, served to oil the wheels of the agricultural
economy and the process of tribute extraction. Sometimes, this
sawthecreation oflinks betweencommercial investment, money-
lending and collection of taxes. In Roman history, such activities
have traditionally been connected with the great companies of
publicani, the tax-farmers who played a prominent role in mid-
die and late republican history. But the phenomenon should by
no means be seen as confined to that period or as restricted to
that scale of activity only.183 In early imperial Pompeii we find
the more humble auctioneer Jucundus also engaging in farming
urban revenues. '^ Documents from second-century Egypt add
a few more tesserae to our fragmentary mosaic. One contract
reveals a member of the local political elite, who was active in
the collection oftaxes, financing the trading trip of a group ofmer-
chants. Another, recent, andbynowfamous, papyms, the so-called
Muziris Papyrus, provides evidence of a loan to finance a cargo
of Oriental luxury goods, presumably intended for the market in
Rome, to the value ofalmost HS 7 million. As Rathbone, among
others,hassuggested,it is difficultnotto seetheinvolvementofthe
imperial aristocracyhere. Investingin a cargoworthseveraltimes
the minimum fortune of a Roman senator was an efficient way of
transferring wealth from the provinces to the imperial capital. 185
In that perspective, the dichotomy between market and state redis-
tribution, which has attracted so much attention over the years,
reveals itselfto bea modernist construct. In theEmpire both activ-
ities, often in direct collaboration, aimed at turning the extracted
tribute into flexible resources which could be disposed of in other
contexts. 186
, "?"? 2007: 3I-54 fora general discussion, butwith a focus on late antique material.
1!<4Cf. the tablets from hisarchive containing receipts confirming payment byJucundus
to the city of what is due on these account: CILIV, suppl. [, pp. 382-404 (tablets nos.
138-151) with Andreau 1974: 53-71.
1 Rathbo"e 2003: 215-225. Foranalysis ofthedocuments, seeCasson 1986;1990;and
Rathbone 2001.
De Salvo 1992: 69-78 is a good example of how meaningless the distinction between
freetradeandstateredistribution canbe.Theboatscarrying thegrainarriving inRome
mayhavebeenfreebuttheircargoesweredominatedbystateshipments.
121
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
Growth and/or tributary concentration
Sucha system of markets developing aroundpolitical exploitation
is what Weber referred to as political capitalism or even "Raub
Kapitalismus" (robber capitalism). 187 According to the standard
view, this is an arbitrary and disruptive sort of economic activity.
But there is an "error" in perspective here. Clearly the type of
economic exploitation which I have described in such detail above
does not fit the doctrines of liberal, legally ordered economics
well. 188 However, while it did not offer the stability and security
of economic assets which we have come to expect today, the close
alliance between empire and aristocratic groups did nonetheless
constitute a solid and enduring kind of order. In a closely argued
critique of the sociological tradition on tributary empires, Michael
Mann has suggested that the alliance between state and aristocracy
was socially productive; their mutual collaboration produced a
synergistic effect which increased extraction and the agricultural
surplus and therefore allowed greater scope for the development
of manufacture and trade. '89
This brings us to the question of growth. But first some clarifi-
cation of the meaning of the term is called for. Growth is far from
being an unambiguous concept. Clearly, from a modern perspec-
tive the most interesting question to pursue is whether the economy
of the Roman Empire experienced significant growth in per capita
incomes. 190 Some have argued along these lines, albeit with con-
siderable differences in degree. 191 Central to this argument has
been the identification of some gradual developments in technol-
ogy, suchasgrowthin thetonnageofships,reductionofthe weight
of amphorae in relation to the volume of their contents, investment
'87 Lo Cascio 1991; Whittaker 1994; Schiavone 2000, chapter 3 have a clear understanding
of the submission of the market to tributary forces. Love 1991: 223-245 on Weber's
political capitalism.
188 Jones 1988: 135 notes that it is difficultfrom a modem perspective to understandhow
markets could work under those circumstances.
189 Mann 1986: 167-174 and chapter 9.
190 Millett 2001 discusses modern notions of growth and the ancienteconomy.
19' Compare the cautious account ot'Hopkins 1995-6 with Mattingly 1996 and Hitchner
2005.
122
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
in biggerolive presses and the like. '92 There may also havebeen
some pockets of rising agricultural prosperity where producers
were able to take advantage ofexpanding markets to increase spe-
cialisation and improve productivity. Such processes have been
suggested for Rome's Italian hinterland (Etruria-Campania)dur-
ing the late Republic and early Empire. '93 Yet it is difficult to reach
an assessmentof the impact ofthose and similardevelopments in
terms of increasing overall productivity in the Roman economy at
large.Mostsuchdevelopments,I suspect,will haveimprovedeco-
nomic performance only marginally. 194 Elsewhere I have argued
that the greater efficiency of African as compared to Spanish oil
amphorae would have lowered the sales price of oil by less than
5%. 195
Ultimately, in the absence of statistics, the measurement of per
capi ta growth will remain beyond our grasp. This should be no
cause for sleepless nights. Per capita growth is in any case a very
problematical concept to apply to an economy which was heav-
ily dominated by pre-modern agricultural production. The classic
illustration of this is economic developments in Europe following
the Black Death. In the aftermath of the pandemic, aggregate pro-
duction plummeted while the decimated population experienced
a strong growth in per capita income as more and better land
became available to the individual peasant. '96 In traditional agri-
culture, increased production is often only achieved at the price of
falling marginal returns. Essentially the peasant is faced with the
choiceofworkingexistinglandharderorexpandingthecultivated
area by bringing more marginal lands under the plough. Either
way the increase in production tends to require a proportionately
bigger increase in the amount of labour. Production expands, but
192 Hopkins 19836;PeacockandWilliams 1986:51-53;Greene2000;Wilson2002.
193 On this, see Scheidel 20073 (building on Morley 1996) and 2007b.
94 Saller 2002 for some cautionary observations; moderate growth can easily explain
developments.
ly5 Bang 2002a,n. 69. Cf. Lelener 2006.
19f> Livi-Bacci [992: 47-54; more generally Braudet 1981: 32-34, pace Persson 1988 (in
the long ran no pre-industrial agrarian economy can really escape this constraint, cf.
Wrigley 1988.Stagnationshouldnot primsirilybeput downto a result of exogenous
shocks, as Persson 1988: 131-135 also has to concede). See Scheidel 2002 for an
attempt to understand theeconomic effects ofthe Antonine Plague affecting the Roman
world in the l6os AD in terms parallel to the Black Death.
123
 THE ROMAN EMPIRE AND COMPARATIVE STUDY
productivity declines. The income of the peasant grows but it is
by no means clear that his living conditions improve correspond-
ingly. 197 Obviously this puts very severe limits on how much the
production of the individual peasant households can be increased.
That being so, the total command over or availability of human
labourisprobably a muchbetterkeytounlocking thesecretsofthe
economic performance oftheEmpire thanthe notion ofsubstantial
and sustainedincreasesin productivity in the economy at large.
Tobeginwith,theincreasedpowerofthetribute-cum-rentnexus
may have succeeded in forcing peasants to work harder. This can
be seen as a process in which the tributary system mobilised the
resources ofmral underemployment. The scope for such intensifi-
cation varied considerably according to local conditions. It was
greatest and most visible in the western provinces where sur-
plus extraction was brought closer to the greater level of inten-
sityrequiredto sustainMediterranean civilisation. '98Butboththe
growth of slavery in the wake of the Roman conquests and the
development which can be followed in the legal sources, where
a distinction between honestiores and humiliores from the early
second century AD gradually overtakes the distinction between
Roman citizens and aliens, may indicate a general intensification
in the exploitation of ilpopulo minuto. 199
It is questionable, however, whether this trend towards inten-
sified cultivation ever materialised as a sustained increase m per
capita production. There were countervailing forces. Completely
stationary populations are rare. It now seems generally agreed that
therewasa modestriseinthepopulationfigureduringthefirsttwo
centuriesAD. Themost plausiblerecentestimateseesthepopula-
tion growing from around 45 million in the ageofAugustus tojust
over 60 million in the middle of the second century AD before the
onsetoftheAntonineplague.200As a consequence,existingfields
197 Boserup 1965;Sahlins 1972:Chayanov 1986,chapter2.
lt)8 Woolf 2001. See Hopkins 2000: 257-259 for some model calculations showing how
rents and taxes increase the area of cultivable land neededby the peasant household
to sustain itself. Insteadof more land, peasantsmay haveto work existingfieldsmore
intensively.
199 N0rlund 1920;Garnsey 1970;Finley 1980,chapter4 on this. GarnseyandHumfress
2001, chapter5 for a nuanced discussion.
200 Frier200oa: 8 [2-8 it.
124
AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE
would have had to be shared by more producers and the area of cul-
tivation to be extended into new zones, bringing less fertile, more
marginal lands under cultivation. 201 This would have pulled in the
opposite direction, producing a tendency towards falling per capita
production, as individualpeasantfamilies wouldhavehadto work
harder simply to achieve the same outcome. Even so, in aggregate
terms the economy would still have continued to expand as long
as the new producers were able to grow a surplus. 202 Obviously, it
will never be possible to work out the combined economic effect
of these two countervailing forces of agricultural intensification
in exact terms. But we may gain a clearer understanding of the
problem through a hypothetical experiment. If we assume that the
politically induced increase in household production would have
been nullified by a fall in productivity caused by the need to culti-
vate more marginal lands, thus preserving per capita production at
an unchanged level, we are still looking at an aggregate expansion
of the economic surplus by around one-third between the time of
Augustus and the early years of Marcus Aurelius. In other words,
actual per capita growth is not necessary to explain much, if not
most, of the increased level of activity that archaeologists have
documented for the early Empire. 203
This points to by far the most important aspect of the devel-
opments presented by the Roman economy: sheer mass. The for-
mation of a Mediterranean empire need not necessarily have led
to a substantial growth in per capita production, at most only a
moderate increase. 4 But the Empire brought together the pro-
ductive capacity of many more people under the same tributary
101 Developments in Africa are a good example, Barker and Mattingly 1996, Dietz,
Sebai and Ben Hassen 1995, 0rsted et at. 2000. So is the Syrian limestone mas-
sif, Pollard 2000: 201-236. Both expanded the agricultural frontier into marginal
zones.
Pace Hitchner 2005: 222, n. 30 who holds that the Roman Empire stayed clear of the
constraint of falling marginal returns. But he has misunderstood the concept. It does
not imply thatproductionfalls off, as hethinks, only that increasedproductionhappens
at lower productivity.
3 Mattingly and Salmon 2001: 3-14 in general emphasise the difference in scale of
individual economic phenomena as a central distinction between Greek and Roman
economic history. Greene 1986; 2000, Wilson 2002 assembles much of the archaeo-
logical evidence for the increased scale of Roman economic activity.
204 SeeSaller2002 for a moderategrowthscenario.
125
 THE ROMAN EMPIRE AND COMPARATIVE STUDY AN AGRARIAN EMPIRE BETWEEN MARKET AND TRIBUTE

umbrella.205 That was the decisive feature of the imperial econ-
omy. It enabled it to mobilise, amass and concentrate resources
from a much wider area than its predecessors. As a result, the
dimensions of the political economy changed dramatically onto a
vastly different scale of activity. 20 The clearest illustration of this
is the formation of very large, individual market concentrations.
The imperial capital Rome is, of course, the leading example. With
a population generally believed to have reached the million mark,
it constituted a concentration of demand not to be equalled in
Europe again until the Industrial Revolution. However impressive,
Rome should not monopolise our attention. Compared with the
urban network of early modern Europe, the system of cities which
made up the Empire stands out by the more dominant position
occupied by large urban concentrations (over 100, 000). Alexan-
dria, Antioch and Carthage taken together are generally believed
to havecomprisedperhaps anothermillion inhabitants.207 In addi-
tion, there are almost a dozen other putative candidates for the over
100, 000 category. As late as 1700 Europe still did not decisively
exceed that number, even though it had a larger population base.
Instead it had many more medium-sized cities (over 10,000). In
l8oo, between 300 and 400 had reached that level. The numerous,
but quite modest, cities in the Roman world notwithstanding, a
similarly dense layer of medium-sized cities is hard to identify. In
other words, the tributary system maintained a much steeper urban
hierarchy, concentrating resources from a very broad base into a
relatively restricted numberof very largecities.208
If the Mughal comparison is anything to go by, the stronger
concentration of surplus consumption would have increased the
scale of individual resource flows, creating some large "centralised
trades" and "arterial routes". 2°9 Such a scenario seems equally rel-
evant for the Roman Empire. The supplies reachingRome are an
obviouscaseinpoint. Butthepattern isalsofoundintheprovinces.
Gaul, for instance, presents an image of strongly concentrated
trades closely linked with the imperial system. Here the Rhone-
Sa6ne-Rhine axis stands out as the dominant artery, serving as the
supply route for the imperial legions stationed on the Gennanic
border. This gaverise to a system wheretheneedfor servicing the
state transports favoured the development of a private commer-
cial infrastructure andtrade. But outside this interregional arterial
channel, the exchange economy seems to have been operating in
rather more local or regional cells. 210
in Syria was given a very large hinterland to dominate and administer. Alexandria.
too, benefitedenormously from its dominantpolitical position in Egypt. The same
goesforCarthage,whichreceivedanextensivechora(territory) to drawits resources
from.Again,theeffectsofempireweremanifest,notonlyindirectexpenditureofthe
imperial patrimonium, butalso in thefashioning andmaintenance ofprovincial elites.
2"t) Bayly 1983: 65 and 159 and chapter 3.
210 Middleton 1983;Whittakerl994a;Jacobsen1995.

205 Livi-Bacci 1992: 109 following Boserup 1981 sees the aggregateeffect of large popu-
lations as more important for the formation of large urban concentrations than marginal
variations in per capita income under pre-industrial conditions.
20 Cf.Wickham2005arguingthatthekeydifferencebetweentheeconomiesoftheRoman
and post-Roman West is the gradual collapse of taxation; this significantly reduced the
scale of economic activity.
207 Hurst 1993 reminds us not to inflate ancient urban populations. Carthage, though still
large, may have been smaller than is often thought.
208 Woolf 1997points inthesamedirection butwarnsagainstseeingtheempire asprimarily
responsible for the strongly pronounced hierarchy in the East. The urban system had
alreadyexistedbeforeimperial rule. Thereforehe points to tradeandcultural exchanges
asequally importantin shapingthe system. But, ashealso recognises(p. 12),thegrowth
of large cities in the eastern Mediterranean was closely linked with the formation of
larger states. Moreover, Roman rule maintained this hierarchy by underwriting the
strength of the aristocracies residing in the most prominent urban centres. Antioch

126
127
 PART II

IMPERIAL BAZAAR

The bazaar is more than a place set aside where people are permitted to come
each day to deceive one another, and more, too. than one more demonstration of
the truth that, under whatever skies, men prefer to buy cheap and sell dear... It
is a distinctive system of social relationships centering around the production
and consumption of goods and services (i.e. a particularkind ofeconomy), and it
deservesanalysisassuch. .. Bazaar,thatPersianwordofuncertainoriginswhich
hascome to stand in Englishfor the oriental market, thus becomes, like the word
market itself, as much an analytic idea as the name of an institution.
Clifford Geertz, Meaning and Order in Moroccan Society

That that trade was not of a "bourgeois commercial" sort has already been
indicated... The financiers of trade were often princes, religious dignitaries,
and nobles. The merchant gentlemen, as popolo grosso, led a seigneurial way
of life when compared to the popolo minuto, the masses of craftsmen and
"actual traders". Furthermore, the princes and lords were themselves often mer-
chants ... Their trade could be occasional or regular. The former could result in
sudden obstruction of markets, forced purchases, temporary monopolisation.
J. C. van Leur. Indonesian Trade and Societv
 3

A ROUGH TRADING WORLD: OPAQUE,

VOLATILE AND DISCONTINUOUSLY
CONNECTED MARKETS

Weknowhowvariedthepricesofthingsarebetweenindividualcitiesandregions.
Gaius, Digesta 13.4.3

In short, whatever is obstructing the development of a modern economy out of

the general backgroundof the bazaareconomy, it is not lack of a "business-like
orientation on the part of the pasaar traders .. . What the bazaareconomy lacks
is not elbow room but organisation, not freedom but form.
Geertz, Peddlers and Princes, p. 47

Some thirty years ago a sanctuary for a hitherto almost unknown

ancient goddess, Nehalenia,was found on the Dutch coast, in the
former Roman province of Germania Inferior. The most remark-
able part of the find was the discovery of a fairly large body of
votive inscriptions put up mainly, it seems, by sailors and mer-
chants thanking Nehalenia for their safe return: "To the goddess
Nehalenia for the good preservation of his merchandise Marcus
SecundiniusSilvanus, trader in pottery to the province of Britain,
kept his promise freely anddutifully. '" Even in theirconcise,mat-
ter of fact appearance the inscriptions provide us with a chilling
glimpse into the world ofmerchants andothertravellers in ancient
times. At best, travel was a slow and arduous trial; complaints over
long delays, bumpy roads and never-endingnauseaare legion. At
worst it was a dangerous undertaking for which you might have
to pay with your life; wreckage, illness or assault by pirates and
highwaymenwere always lurkingjust aroundthe corner. In such
' AE, 1973, 370: "Deae N[e]haleniae | ob merces recte conservatas M(arcus)
Secund(inius?) Silvanus | negotiator cretariu[s] | Britannicianu[s] | v(otum) s(olvit)
l(ibens) m(erito). " The inscriptions can be found in AE 1973, 362-380 and 1975, 641-
656 with furtherexamplesin 1980and 1983.Seealso StuartandBogaers 1971 and200l,
two catalogues from exhibitionspresenting the find.
2 Successfullycompletedtravels wereaccomplishmentstotakenoticeof. Hencethefuneral
inscription (5/G3 1229), mentioning the seventy-two safejourneys madeby a traderfrom

i3i
 IMPERIAL BAZAAR
circumstances one could never be too sure. The protection of the
gods was enlisted as a matter of routine. Indeed, the superstitions of
mariners have at all times been legendary and have provided a rich
material for novelists across the centuries. Think, for instance, of
Melville's immortal description of an American whalingcommu-
nity in Moby Dick, not to mention the more comical chord struck
by Petronius exploiting the prohibition against cutting one's hair
at sea to put his hard-tested (anti)heroes in yet another ridiculous
pickle.3
This secondpart ofthe bookwill explore the conditionsoftrade
in the RomanEmpire. The first part identifiedthe broadcharacter-
istics of the political economy and the main driving forces behind
interregional trade. The inquiry now proceedsto examinehowthe
formation of a tributary empire affected the institutional set-up
for trade. Further comparisons drawn from Mughal India will be
employed to suggest new ways of thinking about Roman traders.
A central concern will be to integrate the risks and uncertainties
of ancient commerce into our general understanding of market
institutions. This third chapter will deal with the basic character of
markets and conditions of trade in the Roman Empire. Emphasis
will be placed on the limits to market integration and the volatil-
ity of commercial links. A middle section attempts to deepen the
image of market fragmentation through a very tentative analy-
sis of the few and scattered price records surviving from Roman
Egypt. The chapter then moves on to consider evidence of enduring
cultural and institutional fragmentation within the Roman trading
world, in spite ofthegreaterlegal andmonetary coherencebrought
about by the imperial government. At the end, the bazaar, with its
many irregularities and uncertainties, is suggested as a model for
the functioning of Roman markets.
Hierapolis, Flavius Zeuxis, around Cape Malea to Italy. Andre and Baslez 1993: 437-
447 ancl483-526 is a convenient selection of the material pertaining to the insecurite
chronique of travel in the Greco-Roman world. See also Millar 1981 using Apuleius'
Metamorphoses to give a vivid impression of the hassles and uncertainties of ancient
travel.
3 Petr. Sat. 104-105.
132
A ROUGH TRADING WORLD
Transport
By modem standards, the means of transport were both slow and
very much at the mercy of the elements. 4 The latter was more
true of travel by water than by land. However, even on land the
weathercouldaffecttravelling conditions significantly. Duringthe
rainy seasons, rivers and lakes might rise above their usual bound-
aries and cause roads to flood, flush away bridges and seriously
impede crossing. Roads, either not built to the best of Roman
standards or not properly maintained, might turn into bottomless
mud holes. In the colder regions of the empire winter could see
roads transformed into slippery ice patches whereas burning sun
in some of the hotter areas would often have forced travellers to
go by the cool but also more frightening night rather than during
daytime. Employed as pack animals or used as traction power for
carts, donkeys, mules, oxen and camels (horses only to a lesser
extent) constituted the basis of Roman land transport. That made
it, on average,considerablyslowerthan waterbornetransport, but
above all a lot more expensive, both relatively and absolutely. The
animals needed considerable amounts of fodder to keep going.s
Working from Diocletian's price edict, A. H. M. Jones suggested
that, roughly speaking, transport on land by wagon increased the
price ofwheat by 50% for every 100miles whereas thecorrespond-
ing figure for seatransport was only i . 3%. One should not pay too
much attention to the exact figures. The relative proportions ofthe
price of land and watertransport would have varied considerably
according to local circumstances. As a rule of thumb, however,
the relationship, which seems to receive comparative confirma-
tion, between the price of sea, riverine and land transport can be
estimated as I : 5 : 5o.6
Braydel 1972: 246-311 remains one of the best descriptions ofpre-modem transport
in the Mediterranean. What follows draws heavily on his analysts. For ancient travel.
collect much
working from
thepriceetlict ofDiocletian.withtheadjustmentsofDuncan-Jones 1982:366-369.
Seefurther Hopkins2000:263, n. 18and 19835, i02-[05. dark andHaswell'1970,
chapter 12 lists some comparative figures of transport costs.
133
 IMPERIAL BAZAAR
As Braudel was one of the first to observe, and as many ancient
historians now agree, that did not reduce land transport to total
insignificance. 7 The high price on land transport was more of an
impediment to low-value bulky goods, such as grain, than low-
bulk, highly priced luxuries where transport costs constituted a
much smaller percentage of the total price. In special circum-
stances land transport might even compete successfully with water.
During the first century AD the caravan route bringing frankin-
cense, of the best and purest crop, from southern Arabia to Gaza
on the Mediterranean seaboard, continued in spite of increasing
navigation in the region. The caravanscould leave in the beginning
of the autumn while ships had to await the arrival of the north-east
monsoon. This gave the caravans a time advantage. They could
normally expect to arrive in Gaza as the ships left South Arabia.
The caravan merchants were thus able to bring their goods first to
the market. This would quite possibly have been sufficient com-
pensationfortheirhighertransportcosts. Evenmorebulkygoods
such as pottery can at times be found to have travelled consider-
able distances overland in large quantities. The famed production
of red-glossed fine ware in La Graufesenque (floruit first century
AD) in southern France had no nearby waterway. Pots, cups and
bowls would have travelled in the thousands on the roads leading
away from the production centre each year. 9 Even so, in general
there is no way of escaping the fact of the primacy of water for
transport over longer distances. The economic logic dictated this
and this logic was not lost on the ancients either. It is noticeable
how the elder Pliny in his geographical description of the Empire,
when assessing the potential resources of an area, repeatedly takes
time to observe whether a river is navigable or not; that would
determine how easily the surplus produce could be mobilised.
In terms of the value of transported goods, Braudel estimated that
in the sixteenth century the relation of water to land transport was
7 See Braudel [972: 284-295. See Laurence 1998 and Harden and Purcell 2000,
chapter 9 for two recent discussions stressing the vitality of land transport.
Young200i: l04-Io6.
9 See Whittaker 2002; see Goody and Whittaker 2001; for proto-industry in southern
France. See further, Vernhet 1991.
10 E.g. Plin. Nat. Ill, [2: "Singilis fluvius... ab ea navigabilis" and 21: "Hiberus amnis
navigabili commercio dives".
134
A ROUGH TRADING WORLD
something like three to one in the Mediterranean. We cannot know
whetherit would havebeen exactly the same in Roman times. But
it does provide us with a useful pointer.
So it is almost true of the pre-industrial world that land divides
whereas water unites - but only almost. Sea transport was far less
expensive and often a necessary requirement for large bulk trans-
fers. However, it was also much more sensitive towardschanging
weatherconditions. On the micro level, we see it when a ship en
route had to go into the nearest harbour and wait, in unfavourable
seasons perhaps even for weeks, until the storm blew itself out
andthe seacalmeddownagain.If the ship foundno ready shelter
and still survived the adverse conditions it might find itself blown
off course, occasionally to such an extent that the only practical
response was to change its destination. On the other hand, some-
times there would be too little wind and the ship would just have
been drifting with the currents until the wind rose again, orthe crew
would have had to row. " On the macro level, we encounter the
greatdependencyon the weatheras an organisationof navigation
around the rhythm of the changing seasons. During wintertime
the risk of sailing increased substantially in the Mediterranean
and navigation came almost to a standstill or at least the inten-
sity of shipping was much reduced. The Romans talked of a mare
clausum at least between November and March/April, with most
navigation probably taking place safely within the margins of the
summer period. 12
All this workedto make travel times on the seavery uncertain
andhighly fluctuating. In his NaturalHistory, Pliny informs us of
some record crossings, such as from Puteoli to Alexandriain nine
Heliodorus Aethiopica V offers many illustrations of the changing fortunes of life at
sea, especially in the more unfavourableseasons(chapter23 for roving). See Synesius
Litt. 129 for an example of a ship being blown so much off course that it changes its
destination to Alexandria.
12 SeeRouge 1966:32-33 formareclausum. Itisclear,however, thatsomesailingdidtake
placeduringthewinterseason.IntheeasternMediterranean,forinstance,sailingduring
thewinterbetweenRhodesandEgypt is attested for anearlierperiod,cf. Demosthenes
Against Dionysoclorus 30. Heliodorus Aethiopica V, chapter 18 points in the same
direction. But here it is also observed that it took special skills and, in the end, even
navigatorspossessingthosemightchooseto winterin a harbour,at leastforpart ofthe
bad season.
i35
 IMPERIAL BAZAAR
days. 13 But that was unusual. If we are to judge from the speed
of imperialcommunicationsomethinglike threeorfourtimesthat
numberofdayswouldhavebeenmorecommon,butwithconsider-
able individual variations andwith some trips taking substantially
longer. 14 The general situation does not seem to have been very
different from the Mediterranean in the sixteenth century. For that
period, Fernand Braudel estimated the length ofa seavoyage from
east to west to have been in the order of two to three months
and often even more, with substantial additions to be made for
penetration beyond the coastal regions. There had of course been
some technological modifications since antiquity, but they were
marginal improvements. The pattern of relatively slow, irregular
and for periods even discontinuous transport and communication
lines, had still to be broken. The same goes for tonnage. Most
shipsin bothperiods wouldhavebeenrelatively small, below TOO
tonnes and more often than not closer to 50 tonnes. Bigger ships
were certainly built and used, especially on routes between the
largest cities, but they were far from the norm. 15
Volatile and opaque markets
Mysteriously this basic continuity over many centuries has been
taken by some classicists asconfirmation that communication and
transport did not impose strong limits on the economic perfor-
mance and workings of markets in (Roman) antiquity. To be sure,
a lot of movement of commodities on sea and on land did occur.
13 Plin. Nat. XIX, 3. '4 Duncan-Jones 1990, chapter 1.
15 Ingeneral Braudel 1972: 295-311 and354-375 remains fundamental on theseproblems
and la longue duree in Mediterranean transport, in spite of the ambitions of Harden
and Purcell 2000 to overtake him. Parker 1992: 26, based on a catalogue of some
1, 200 ancient shipwrecks, concludes that most ships were a good bit below 100 tonnes
capacity. P. Bingen 77, treated by the editor Heilporn and by Rathbone 2003: 226,
contains information about some eleven ships arriving in an Egyptian port, probably
AlexandriaduringthesecondhalfofthesecondcenturyAD.Twoofthese,onea Roman
grain freighter, the other used to carry pine trunks, were well above 100 tonnes. The
majority, however, hada tonnage between 30and75 tonnes. Rouge 1966:47-80 argued,
before the rise of underwater archaeology, for the comparable dimensions of shipping
in the Roman Empire and the early modem Mediterranean. His general conclusion is
convincing, but he tends to overestimate the size ofthe ships. He would have done well
to consulthisgreatcompatriot, Braudel,whohadalreadywrittenon this.
136
A ROUGH TRADING WORLD
But the optimism of Harden andPurcell, reflected in the deceptive
catchphraseconnectivity, is misplaced. The proper point to make
is rather the opposite: transport remained a serious challenge or
even obstacle to economic integration for most of pre-industrial
history. 16Individual markets were not easily tied together so asto
behavein unison.Withslow,irregularandperiodicallydiscontinu-
ous transport and communication it was simply very difficult, on a
regional basis, on the continental scale ofthe Roman Empire quite
impossible, to match demand and supply across individual market-
places in an even, steady and predictable way. 17 Rather, resources
would normally have come in uneven, not easily predictable,
clusters. The problem is conveniently illustrated from Apuleius'
GoldenAss by the story told by oneofits characters,a small trader.
Havinghadnews of the availability of some goocl-quality cheese
at a very attractive price in a neighbouring area, he rushed to the
place only to find that everything had already been sold. His efforts
had been in vain; the market had already been cleared. 18
Acting on wrong or out-of-date information, which did not cor-
respond to the actual situation in the marketplace on the time of
arrival, was an ever-looming risk for the pre-industrial merchant.
It was not just that the goods might have cleared by the time of
his arrival. Too many competitors might havepickedup the same
news, flocked to the market and causedthe low price to increase
violently so as to endanger the profitability of making the trip in
the first place. On the other hand, high prices might attract too
many sellers trying to out-compete each other; or, alternatively,
buyers might fail to show, causing a shortage in demand and the
market to plummet well below normal levels. '9 The latter seems
to describe the situation in the example above from the point of
view of the cheeseseller. He had to part with his goods at a very
low price to a single well-positioned merchant whojust happened
'b Braudel1972:375-379 conveys a goodsenseoftheproblems andfrequent breakdowns
in economicintegration. SeealsoShaw2001:423-425 for a relatedcritique.
17 See Erdkamp 2005, chapter4 forananalysis of the Roman grain market stressing low
[8
carry-overofstocks in theoperationofmarkets through time andacrossdistances.
Apul. Met. I, 5.
19 SeeforexampleP. OsloII,63(thirdcentury).Thisisanestateletterinwhichanagent
reports that he cannot sell his goods at the price at which he has been instructed to sell
them. The market is much lower.
137
 IMPERIAL BAZAAR
to be in the area. Another illustration can be found in one of the
letters ot'Cassiodorus. The state needed to make some requisitions
from an area with plentiful stocks. But it was out of season and
there were few merchants in the area. Therefore demand was low
andpeople were forcedto accepta very unattractivepricefortheir
goods. 20
Closely connected to the information problems were logistical
difficulties. A merchant might act on correct information and still
seehisplansfall through.Hisshipmentofgoodsmightbeseriously
delayedenroute. Or,alternatively, hemighthaveto leavea market
before the goods he wanted to buy hadarrived, either dueto delayed
cargoes or becausefair weathermade it expedientor the close of
the trading season made it imperative for him to set sail. 21 It would
also, occasionally, be difficult to find a ship which could bring
orders to an agent in another harbour at the required time. Equally
it could be a problem to find cargo space for hire when needed.
Surplus of shipping in an area with nothing to export and shortage
in a neighbouring region with stocks ready for export were not
unheard of.22 In other words, the pre-modern merchant had to act
in a highly uncertain market situation, governed by the rhythm
of the seasons, where it was difficult to predict or estimate the
amounts of goods brought to the individual marketplace and the
number and buying capacity of his competitors. Such markets
would normally be quite prone to violent fluctuations in the short
term. Price changes of say 10, 20 or 30% within a few months,
weeks or sometimes even between days were the order of the day,
making it a very risky business environment. Under the ruling
regime of communication and transport it is difficult to imagine
20 Cassioclorus Variae XII, 22. 2. Rouge 1966 diligently gathered much late material show-
ing the irregularities of Roman trading, especially pp. 416-417 for more examples.
2' Seasonalityoftrading:PhiloLeg. Gai. 15;Sirk.s 2002(on Dig.45. 1. 1 22. 1).TPSiilp.46,
53 and79 (asexplainedby theeditorCamodeca)showa loanbeingadvancedin March
to a Puteolan grain trader on the security of a consignment of grain. Had the loan not
been repaid by the Icfes of May, the lender was entitled to auction off the grain just in
time to beat the arrival of the Alexandrian grain fleet whose fresh supplies would cause
prices to fall in the market. See also Cic. Fain. XIV, 4, 5 (ship departing Brundisium
before arrival of letters in order not to lose fair weather).
22 Cassiodorus Variae XII, 24. The letters of Synesius offer some striking illustrations.
No. 129 mentions letters which have gone astray and have been delayed many months.
Nos. 134 and 148 describe problems about finding cargo space and ships heading to the
required destination.
138
A ROUGH TRADING WORLD
how the hidden hand could ensure that such fluctuations would
find a strong and ready response in other markets that would be
able to iron out the differences and create a more stable situation -
except in a very imperfect, almosthaphazard,way.
Ceteris paribus, the volume of transactions alone was not
enough to establish stable markets with closely correlated move-
ments of prices. In the short run there would normally be only a
very weak link between price developments in particular demand
markets on the one hand and the quantities of a given commod-
ity sent from the various individual supply markets on the other
hand. 23 The result was a trading world characterised by chronic
bottlenecks, imbalances and asymmetries in the supply of avail-
able information and goods. A humorous letter of Pliny sums up
thesituationofdearthinoneregionandglutwithcollapsingprices
not too far away:
My Tuscan fields, I hear, have been wiped out by hail, and in the Transpadane
there's a bumper crop but prices havecollapsed in consequence. Only my Lau-
rentineyieldsmea return.OfcourseI havenothingtherebeyondhouse,garden
andsandybeachhardby; still for me it is the sole sourceofprofit. For'that's
where I do most of my writing. I don't cultivate the land- for I havenone- but
myself in my studies. 24
In recent years economists have become increasingly interested
in the consequences of such market imperfections and asymme-
tries. So far their efforts have been rewarded with several Nobel
prizes sincethe 1990s.25 Therecipients arguethattraditionaleco-
nomics have overestimated the tendency of markets to become
integrated and move in unison. A classic illustration of this is
the huge gap between the comparatively moderate interest rates
in Indian cities and the usurious rates that Indian peasants have
often had to pay to the village money lender. For he, unlike the
23 Steensgaard 19733:58;ingeneralpp.22-59foraninci.siveanalysisofmarketscharac-
terisecl by weak integration and short-term fluctuations.
Plin. £p. iy, 6, [-2:"Tusci^randineexcussi,inregioneTranspadanasummaabundantia,
seeIpar vilitas nuntiatur. Solum mihi Laurentinum meum in reditu. Nihil quidem ibl
possideo praetor tectum et hortum statimque harenas, solum tamen mihi'in'reditlL
Ibi plurimum scribo, nec agrum quem non habeo sed ipsum me studiis excolo. " The
passageisanalysedbyTchemia 1986: 178.Erdkamp2005: 167-169isspeculativeand
overcomplicates the issue.
25 ForexampleDouglassNorth,GeorgeAkerlof, MichaelSpenceandJosephStiglitz.
i39

IMPERIAL BAZAAR
banks, knows the individual borrowers. He can therefore better
assesstherisk involvedin advancinga loanto a particularpeasant,
the argument goes. Should a middleman attempt to bring thecity
andvillage creditmarkets together, herisksattracting all thebad
payers, due to his lack oflocal knowledge and soon goes out of
business. Consequently the village money lender rests secure in
his local monopoly and the markets remain separate. The lesson
canbeputinmore general terms: market environments withhuge
irregularities, low transparency, great uncertainties and slow and
at times erratic transport, make it expensive and difficult for eco-
nomic agents to transact andmove goodsfrom market to market
When transaction costs are high, it cannot be taken for granted
that markets will tend towards the same equilibrium. 26 Such con-
siderations areofobvious interest to students ofpre-modem trade.
One crucial implication for our understanding of the Roman (as
ofmost ofthepre-industrial) trading world, is that any notion of
the empire as constituting one generalised market, where prices
converge around a common equilibrium price across the many
individual markets, cannot easily be maintained. 27
In contrastto marketin the abstract,generalised,modem sense,
it iscustomary to speakofconcrete, individualmarketplaces exist-
ing in isolation. But here it is important to keep in mind that we
are dealing with two ideal-types at each end of a broad spectrum
of varying"degrees of integration. 28 The purely concrete market-
type mightreasonably catchthesituation in someRoman inland
towns far from waterways or the most frequented roads. Gre-
gory of Nazianzus, explains the plight of food shortage in the
mland city of Caesarea in Cappadocia: "Coastal cities support
such shortages without much difficulty, as they can dispose of
26 Forsomepioneringstudiesofthenewperspective,seethenowclassicpaperbyAkerlof
(1970),alsoRotschild 1973andNorth 1990,especiallypages60-80.
^ Temin 2001: 179isanexampleofthetendency oftraditional economics to playdown
the'effe'cts'ofmarketimperfections:theymightin theshortruncreatennbalanresbut
theycould not prevent an equilibrium from materialising across the Roman Empire
iri'the middle- or long-term perspective. It is precisely this expectation that the new
institutional economics has undermined. _
28 Marshall 1890: 323-330, the founder of neo-classical economics, already made. this
obse'rvation/But~neo-c~iassical economics has never been very interested in exploring
thespaceinbetweenthetwoextremes.Hencethesuccessofinstitutional andtransaction
cost economics today.
140
A ROUGH TRADING WORLD
their own products and receive supplies by sea; for us inland our
surpluses are unprofitable and our scarcities irremediable, as we
have no means of disposing of what we have or of importing what
we lack. "29 But even inland areas were not completely isolated.
Grain, for instance, was certainly transported out of the innerparts
ot'Anatolia to be sold in Mediterranean markets. 30 Nonetheless, it
was a costly and burdensome undertaking from which cities some-
times sought to relieve themselves at a price. As Cicero explains,
inland cities might accept to pay, in lieu of their grain taxes, a lump
monetary sum calculated on the basis of the higher prices prevail-
ing in more busy coastal markets, rather than having to deliver
their taxes in kind on the coast. 31 With cities enjoying close access
to waterways, however, the concrete market-type begins to require
considerable modification. Perhaps the situation can be more use-
fully described in terms of concrete connections. It is the density
and stability of those that we need to look at.
Most cities in the Mediterranean world were small (less than
10, 000 inhabitants). Markets, therefore, were generally small too
and could not easily sustain close permanent connections, espe-
cially because the needs of many cities will have fluctuated from
year to year. This is most marked in agricultural products where the
strong interannual fluctuations in harvests would have made cities
constantly change between the roles of exporter and importer of
small amounts of grain etc., as illustrated by the above quotation
from Gregory. Those markets were normally served by a system
often referred to as cabotage: small merchant ships would more
or less casually tramp along the coast from harbour to harbour in
search of a good bargain. 32 This, combined with the constant shifts
in the needs of markets, did not work in favour of bringing areas
closely together. On the contrary, the loosely linked small cabotage
markets, characterised by impermanent connections, accentuated
the uncertainties and risks of pre-modern trade. The arrival or
non-amval of as little as one or two extra ships during the sailing
9 Gregory ot'Nazianzus In Laudem Basilii 34-35 (Migne, Pair. Gr. 36, 541-544, quoted
from Jones 1964: 844). See further the discussion of Gamsey 1988: 22-24 for a per-
ceptive analysis, pointing out that Gregory probably overstates his case.
3 Mitchell 1993, vol. i (last chapter, analysing the economy).
31 Cic. II Verr. 3, 191 with Erdkamp 2005: 197-198.
3 Harden and Purcell 2000, chapters 4 and 5 make a great deal out of this.
i4i
 IMPERIAL BAZAAR
season might have had a decisive influence on the price in the
market. Consequently prices would often be highly volatile and
the markets would be dominated by interlocal asymmetries and
imbalances.33
The situation begins to change somewhat when attention is
turned tothe small groupofvery largecities/markets intheEmpire:
Rome, Alexandria, Antioch, Carthage, Ephesos and a few others.
In somecommodities,suchasthemosthighlypricedluxuries,mar-
kets will still have been very narrow. This is what Cicero alleged
had been the experience of the Roman knight Rabirius Postumus,
whohad imported a valuable Egyptian cargo of"paper, linen and
glass" to the imperial capital. But the promise of a rich profit had
turnedout to be "treacherousanddeceptive",becausemany other
ships had arrived in Rome carrying the same goods and caused
prices to fall. 34 But such luxury markets were nonetheless very
much larger and represented a far more stable concentration of
demand than could be mustered in the smaller towns. This is even
moretrue inthecaseofstaplefoods.Therequirementsofthelarge
citieswereofa sizethatcouldnotbesuppliedinthesmallscale,ad
hocmannerofcabotage.Totakethemostextremeexample,Rome
would probably have required something like 1,500 shipments of
loo tonnes to supply its needs for grain.35 Clearly, more perma-
nent connections and larger concentrations of merchants, transport
facilities, infrastructure and routes to maintainthese hadto come
into existence in order to serve the needs of such giant markets.
The big cities thus represent several steps up the ladder of inte-
gration. 36From the merchants' perspective, they offered a higher,
33 Pace Harden and Purcell 2000, chapters 5 and9. Gibson and Smout 1995: 270 provide
a clearexposition ofthelowdegreeofeconomic integration createdbyphenomena such
ascabotage,characterisedbyimpermanentconnections,andofthestructuraldifference
from laraermarkets. SeealsoTchernia 1986, 172-193.Goitein 1967,chapters 3 and 4
provides'a very useful description oftheuncertainties ofa worldofsmall Mediterranean
markets in the early Middle Ages. The general conditions would not have been very
different in antiquity.
34 Cic. Pro Rabirio40 ("fallacesquidemet fucosae")withErdkamp2005:177. It does
little that the passage has been emended, best by Mommsen; the general meaning is
quite unequivocal.
35 Gamsey 1983^ 118estimatesRomangrainimports atbetween 150,000and200,000
tonnes wheat.
36 SeeTchernia 1986, 189-193 for a related point. Morley 1996 explores the cre;ition of
more stable connectionsofeconomic integrationcreatedby the demandoftheRoman
142
A ROUGH TRADING WORLD
more steady and predictable demand than other markets. In that
respect, the large consumptive nodal points reduced commercial
risks. But they were far from unproblematic markets to negotiate.
Coming to a large city, without a prior established social network,
carried risks of its own. That theme, however, will be developed
more fully in Chapter 5. For the moment, I am interested in a dif-
ferent aspect; the giant markets were logistical feats that strained
capacities to the full, and supplies were in constant danger of
breakdown or temporary interruption. 37 In that way they remained
nervous and volatile. In a tense situation, even the arrival of as few
as five ships with grain might suffice to ease for the moment the
worries of a riotous urban mob, as one of the letters written by
Sidonius during his tenure of the urban prefecture in fifth-century
Rome demonstrates. 38
Finley's classic illustration of the vulnerable and unstable char-
acter of the larger urban markets was the food crisis in fourth-
century Antioch during the stay of the emperorJulian. Only with
Julian s intervention was the emerging crisis averted, but not even
then without problems, through the importation of grain from
nearby imperial estates and later from Egypt as well. Market net-
works had been unable to solve the problem by responding to the
growing demand and bringing in supplies from surplus areas. 39
The various accounts of the crisis add an extra dimension to our
understandingof the problem. It was not only a question of high
transportation costs. Local power wielders exploited their control
of the agricultural surplus in the region to take advantage of the
shortage and increasethe local market asymmetries by very con-
siderable hoarding in order to drive up prices further and make
market. The interpretation ottered above clashes directly with Harden and Purcell 2000,
e.g. p. i2l, who reduce the larger markets to an outgrowth of cahotageand therefore
tone down their importance. These markets, in fact, represent a qualitative change.
They could not possibly be sustained by the casual small-scale movements of coastal
tramping. Theyneededa farmorepermanentinfrastructure. It is in thisperspectivethat
the major shipping (and land) routes take ona very important role asthemain conveyers
of economic integration.
37 The phenomenon shouldnot be seen as restricted to grain supplies. Thus Suetonius'
biography ofAugustus (42. 1) and Ammianus Marcellinus book 14, chapter 6, I, report
of wine shortages in the Roman market.
3 SidoniusEpistulae[, 10withGarnseyandHumfress2001, 114.
TO Finley >985a: 127. See further Gamsey and Humfress 2001: I20-I2I and Hopkins
.
l978a:46-47 for discussionofthe emperorJulian'sreliefofthe food crisis in 362-3.
143
 IMPERIAL BAZAAR
a killing. The intervention of the emperor provoked resistance in
thesegroups. They continued tokeeptheirgrainfrom themarket in
the city andeven hadthe "audacity" to buy up some ofthe imperial
"relief supplies brought to Antioch. One can easily understand
why more modest merchants hesitated to step in and alleviate this
particular crisis. It wouldhaveamounted to commercial suicidefor
such people, outsiders at that, to challenge the deeply entrenched
localmonopolypositionenjoyedbya verypowerfulgroupoflocal
landowners. 40 The example carries a significant lesson which is
rarely drawn: in the pre-industrial world closer integration of eco-
nomic resources and tighter coordination of markets, in particular,
depended ona handanything buthidden. It wasimpossible without
the intervention of very tangible and overt forms of organisational
power, to tie supply anddemand across locations into a more stable
relationship wheremarkets would be more closely integratedand
behave more uniformly.4
A further clear demonstration of this is providedby Indiain the
aftermathofMughalrule andtheearlybeginningsofBritishhege-
mony. After a disastrous famine in Bengal in the 17708 the British
EastIndiaCompanyattemptedto organisea more stableandinte-
gratedcompany/state-controlledgrainmarketin theprovince. The
scheme was a fiasco and quickly abandoned; the Company had
failed to gain control over the numerous rural supply networks.
Short of that, it was unable to establish a more stable and trans-
parent situation in the grainmarkets in Bengal.The internal grain
market in Bengal, however, was only a secondary concern of the
Company. In areas ofcrucial importance for its operations, such as
opium production andtrade, it didnot tolerate defeat. Here control
waspursuedruthlessly,withwhatevermeansnecessary,untilit was
achieved and the market situation stabilised.42 The Indian exam-
pie is particularly instructive because it so clearly shows the clash
of two different market regimes. There is a growing awareness
amongeconomiststhatmarkets arenever"just"markets;theyhave
40 Julian Misopogon 3680-3705 confirmed by Libanius Oratio [8, 195.
41 Contra Harris 1999, who supposes that the mere existence of a substantial trade in
slaves would have worked to create uniform movements of prices across a large group
of individual, separate markets.
42 Chatterjee 1996, chapter 5 offers a very clear analysis of the British East India Com-
pany's first decades in Bengal.
144
A ROUGH TRADING WORLD
significant individual characteristics shaped by the institutional
fabric which supports them. 43 This is what historians should be
studying.
In dealing with markets in Bengal the EastIndia Company drew
on a set of trading strategies which in the two preceding centuries
had transformed and gradually created a number of more coher-
ent and integrated markets in Europe.44 The various chartered
joint stock companies were themselves a prime expression of this
development. The cmcial characteristic was the ability to create
greater and more long-lived concentrations of trading capital than
hitherto. This enabled the merchants to adopt a longer-term per-
spective of the market and create buffers which could absorb some
of the endemic irregularities of pre-industrial trade. The larger
capital base enabled them to extend control over trading in many
individual markets. Control is here to be taken literally. It was
far from uncommon that military force was employed to achieve
the objective.45 A report submitted to the Dutch East IndiaCom-
pany (VOC) in the late i620s by Francisco Pelsaert, the Dutch
factor in the Mughal capital, Agra, conveys a very clear idea of
this process. 46 VOC's commercial presence in the Agra area was
primarily motivated by the local production of incligo, a highly
desired dye. It bought considerable quantities, to a large extent
financed by selling its primary commodity, spices imported from
various South-East Asian islands. These spices, however, were
alsomarketed in largequantities by the company on the Coroman-
del Coast (the south-east Indian litoral). In too large quantities,
as a matter of fact, according to Pelsaert. The problem was that
43 North 1990. North, in a way, has brought the challenge of'Polanyi into the mainstream
of economic theory.
44 Braudel and Spooner 1967 offer a useful overview of the gradual creation of more
stable, lessfluctuatingandmore uniformpricemovements in Europeanmarketsduring
theseventeenthandeighteenthcenturies.SeeGrangerandElliott 1967andGibsonand
Smout 1995 for two case studies of this process. For the extremely close correlation of
the price of grain in Amsterdam and in Danzig, see De Vries and Woude 1997: 414-419.
45 See Steensgaard 19730,chapter3 andChaudhuri 1985, chapter4 for analysesof the
chartered companies. Glamann 1958 demonstrated how the Dutch East IndiaCompany
did not possess a simple monopoly, but together with the other chartered companies
created and developed a competitive European market in spices. Cf. Pearson 1993, who
providesanotherexampleofthe importanceofthedevelopmentoforganisationsforthe
integration of markets.
46 Pelsaert 1925,chapter5.
M5

IMPERIAL BAZAAR
manyIndianmerchants, willingtoaccepta lowerprofit, imported
the Dutch spices from theCoromandel Coast to the Agra areaand
destabilised the market with falling prices as the result. This did
considerable damage to the Dutch capacity to procure indigo in
the area. Lower prices on spices meant less money to buy indigo.
Ontop ofthat, it becamemore difficult to reachanadvantageous
pricewiththeindigosellers. Seeingthatthefactorwasfinancially
vulnerable they could press him harder and claim a higher price.
Therefore Pelsaert suggested thatthecompany shouldmakebetter
use of its control over the spice trade to cut the supplies from the
Coromandel Coast ofhis Indiancompetitors in Agra. This would,
of course, mean smaller profits in the trade on the Coromandel
Coast. But they would be more than compensated by the greater
income generated by the operations in Agra. What is remarkable
here is the conscious formulation of a strategy to coordinate and
integrate operations in widely separated markets into a coherent
system.
Other aspects of this process included the financial ability to
fund the build-up of extensive stocks, sometimes equal to the
turnover of several years, in order to make trading less dependent
on the erratic fluctuations in supplies. In addition, attempts were
madebymerchantstopushbeyondwholesalemarketsandextend
control right to the door of the producer through the advancement
oflargecreditsor,ifmoreforcefulmethodswererequired,through
military action.Thusthegreaterorganisational powerofmerchants
enabled them to get a much firmer grip on the erratically working
supply networks andre-shape thebasicconditions oftrade. The
result'was thegradual growth ofmore stable andintegrated markets
where it made good sense to coordinate activities in individual
locations closely in order to achieve long-term goals rather thanto
obtain short-term profits. For instance, in spite of achieving a very
high degree of control with European pepper imports, the East
Indiacompanies didnot take advantage ofthe situation to drive up
pricesexcessively andmakea quickmonopolistic profit. Instead,
they preferred to keep the price at a level where the temptation to
challenge theirdominance wouldneverbetoo great. Infact they
were able to operate a gradually narrowing profit margin. While
pricesroseintheproduction areasinIndonesiatheystill managed
146
A ROUGH TRADING WORLD
to lower the price of pepper in Europe throughout the seventeenth
century. 47
Perhaps some will argue that the joint stock companies were
exceptions, not even representative of European trade in the
period. On formalistic grounds that is correct. But in real terms it
would be more precise to say that they were the most articulated
examples of a general trend in institutional developments. When
Dutch supremacy in European trade during the seventeenth century
receives explanation, it is the same general features that are evoked.
Exemplifiedby suchinstitutions asthe Beurs, Korenbeiirs,Wissel-
bank, enormous warehouses and publication of printed biweekly
price notations, it is the unprecedented strength and capacity of
Dutch commerce to control, organise and integrate markets which
repeatedly receive emphasis. 48 Perhaps its biggest achievement
was the level of integration it achieved in the Baltic grain trade.
It is truly astonishing to observe how closely the price in Danzig
followed developments in Amsterdam in the eighteenth century
(Fig. 3.i).49
Strong political support, greater and more permanent concen-
(rations of capital and much improved information, the achieve-
ments of Dutch international capitalism were an expression of a
general process in parts of Europe which saw the position of the
middleman strengthened considerably. The wave of mercantilist
privileges, employed particularly by the rivals of the Dutch, were
aimed precisely at this goal. Later the process was given a fresh
impetus by absolutist governments seeking to push through a pro-
gramme of free trade, in emulation of eighteenth-century Britain.
This strategy was pursued by governments to open up their terri-
tories for trade and loosen the control of local societies over their
own markets. For, as Adam Smith remarked about the grain trade,
47 Steensgaarcl l973a: 146 and 152-153. In general Chaudhuri 1985, chapter4.
4 Apart from Steensgaard 19730 and Chaudhuri 1985, see Klein 1965; Barbour 1966;
Wallerstein 19740; Israel 1989; De Vries and Woude 1997; Lesger 2006 for expo-
sitions of the principles and institutional strength of Dutch trade. See also Braudel
1982 for numerous perceptive comments about the growth of the modern world
economy.
49 See de Vries and Woude 1997: 414-419 for the integration of the grain markets in
Amsterdam and Danzig. Christensen 1941 is the classic analysis of the Dutch Baltic
grain trade and its institutional strengths.
147
 r
IMPERIAL BAZAAR A ROUGH TRADING WORLD

"literally with recourse to the forces of order" in the face of local

(a)
opposition.5'
The question at hand now is whether we should expect
the tributary-fuelled process of market formation, described in
chapter 2, to have causeda similar set of closely integrated mar-
kets to emerge in the Roman Empire. Finley was adamant in his
denial. 52 Not suiprisingly, his critics have seized the opportunity
to argue the opposite. 53 If we turn once more to our Mughal par-
allel, it does not warrantsucha reaction. As the experienceof the
chartered companies shows, markets did not easily adapt to the
more stable forms gradually developing in Europe. Rather they
seem in generalto haveremainedless strongly integrated.Volatile
prices, erratic fluctuations and significant local differences and
asymmetries continued to dominate the bazaars of north India. 54
It wouldbewrong,though,to mistakethisfora signofa primitive
trading world; in certain limited ways the Indian bazaar had at
its disposal tools for transacting business which were, in evolu-
tionary terms, more advanced than those at the command of the
Roman merchant. Most important among those was the hundi, an
Indian form of the bill of exchange. 55 It is now clear to Roman
historians that merchants may to a limited extent have been able
to organise cashless transfers of money, for instance with bankers
^."..v-v-v"r acting as intermediaries. But it is equally clear that a fully fledged
1738/42 1758/62 1778/82 and solidly institutionalised system of bills of exchange did not
Period
materialisein the Romanworld. Otherwise,wewouldexpectit to
Fig. 3. 1 Grain prices in Amsterdam and Danzig in the eighteenth century have left far stronger traces in the historical record than is in fact
(De Vries and Woude 1997: 418) the case (see Chapter5). The performance of the Mughal trading

"no trade deserves more the full protection of the law, and no
trade requires it so much; because no trade is so exposed to pop-
ularodium".50 Imperfectionsandfragmentationdidnot disappear
from European trade. Nevertheless, the result was a gradual sta-
bilisation and tighter integration of markets, both nationally and
internationally, created across large stretches of Europe and the
Atlantic world from the seventeenth till the nineteenth centuries
50 Smith 1976, vol. 1:527.
148
51 Persson1999: 146,ingeneralchapters4 and6. SeefurtherBraudelandSpooner1967.
On local opposition to the process, see Thompson 1993, chapters 4 and 5. The lack
of interestof largeempires in pursuingsuchpoliciesoffavouritism, wouldthuscarry
significantconsequencesfor thedevelopmentoftrade, paceMorley 2007:58.
-"" Finley 19850, "FurtherThoughts", first section.
53 Harris 1993. 1999 and 2000 (the latest piece was written before the two first mentioned
andis more cautiousasbefitsa chapterin The Cambridge Ancient History), Rathbone
1997 (o" Egypt), and Temin 2001.
54 Blake 1987:451-455; Prakash 1991:64-65; Habib [999: 39-62, 68-82and90-102;
and Moosvi 2000 on the "imperfections" of markets in the Mughal Empire. Marshall
1976and Chatterjee 1996complement with material from the late Mughal andpost-
Mughal period.
55 Habib 1972 foranexposition ofthehundisystem in the Mughal Empire.
149
 I
IMPERIAL BAZAAR
world can therefore be seen to offer an upper benchmark for the
Roman experience.
At the same time, it provides us with a useful guide for our
understanding of Roman trade. In spite of the few limited techno-
logical advantages that it held over Rome, it was unable to trans-
form the basic irregularities and volatile conditions of pre-modem
trade. Even after the introduction of the railway and the telegraph
under British colonial rule at the turn of the nineteenth century -
two features which significantly reduced many of the obstacles.
to economic integration - the Indian "bazaar" remained a very
risky and fragile business environment. Many European compa-
nies, therefore, preferred to do business in Indiaoutsidethe sphere
ofthe indigenousmarket; theyneededfargreaterstability andpre-
dictability to satisfy their profit requirements. 5 This goes to show
that the chronic economic imbalances of the bazaar world can-
not be reduced to a simple question of technological deficiencies.
Social institutions played a crucial role. 57 The tributary process
of commercialisation rested on a very different social balance of
power. It did not invest the merchants of the Mughal "bazaar" with
the kind of strong organisational powers and corporate influence
that were necessary to control and conquer the endemic irregu-
larities of pre-industrial trade and production. A telling index of
this is that Indian traders were not normally in a powerful position
vis-a-visthe aristocratic andpolitical elite.58 A small poetic auto-
biography written in the seventeenth century by a Jain merchant,
Banarasidas, provides interesting evidence of this condition from
a merchant's horizon. Aristocrats and imperial authorities in his
story often prey on the readily consumable riches of merchants.
Time and time again traders are taken hostage in order to fill the
empty or just insatiable coffers ofMughal nobles. 59
5f) Ray 1988. Steensgaard 1973;) made the same point for an earlier period.
57 Marshall 1976 offers a pertinent argument in support of this.
58 Marshall 1976; Pearson 1976; Das Gupta 1979; HabibandRaychaudhuri 1982: 183-
202 and 339-41; Das Gupta and Pearson 1987, chapters 4 and 6; Chatterjee 1996;
Subrahmaniam 1999. Pace Subrahmanyam 1995, who remains unconvincing. The few
examples he can point to are marginal and tend rather to reinforce the validity of the
established historiography.
59 Banarasiclas' poem, the Ardha-Kathdnak, written in a dialect of Hindi, has received
two recent translations: a drier, but literally more direct by Sharma 1970 and a more
i5o
A ROUGH TRADING WORLD
The experience of Banarasidas caneasily be matched by mate-
rial from the Roman Empire. In times ofmilitary upheaval we hear
of trading wealth being a convenient target for cash-hungry gener-
alsandsoldiers.InAppian'shistoryofthecivil wars,themerchants
ofPalmyra barely avoid the onslaught ofAntonius by escaping to
the other side of the Euphrates. 60 During the so-called year of
the four emperors, Tacitus tells of a town in northern Italy which,
swelled by merchants coming to a large fair, became a tempting
target for the approaching armies.6' Even more revealing about
the general conditions is the comic anecdote in Apuleius' Golden
Ass about the zealous market official whom the hero encounters
during a visit to the marketplace. 62 Whenthe agoranomos (market
overseer) discovers that the hero has been overchargedfor a fish,
he rushes down to the poor market seller to set him right. Now, the
comic point hinges on the fact that the unfortunate agoranomos
makesa fool ofhimselfbyfailingtogivethepetty hucksterhisjust
deserts and put him in hisproper place, while destroying the fish.
Our hero is left with neither fish nor revenge. Leaving the world
of literary fantasy behind, we have a letter among the Vindolanda
tablets that conveys a similar impression.63 There we encounter
an overseas trader, a homo transmarinus, filing a complaint about
a beating he has received from a centurion. Likewise, aristocrats
could occasionally be found to intervene in the operation of mar-
kets andexploittheirgreatermusclepowerto cornera line oftrade
in order to make a quick profit, as Vespasian reputedly did. 64 In
general, the position of the commercial middleman remained vul-
nerable. Roman authorities often looked with suspicion upon his
activities. Inscribed on a stele in the Piraeus, a set of regulations of
Hadrianattempted to securea plentiful andcheapsupply offishfor
Eleusis, presumably during the Mysteries, by, among other things,
seeking to limit the role of middlemen in the trade; their activities
were denounced as harmful speculation which only served to drive
up prices. 5
poetically evocative intheearly 198os. SeeSnell 2005fora studyofthe literary aspects
of this text.
" AP.P-Bc5'9- (" Tac-Hist-m'30-34. 62 Apul.Met.I,24-25.
.
yi Tab. lWo/. 344. 64 Suet.Vesp. 16. 1. 65 OHver [989, nol77^
l5l
 IMPERIAL BAZAAR
It would be wrong to mistake this greater social vulnerability
of merchants in tributary empires with an absolute powerlessness
or complete lack of political influence. Merchants were far from
always passive victims, asthe letter from Vindolanda also informs
us. They, too, could stand up for themselves or seek the patronage
of powerful people. 66 However, compared to the world of their
early modem European counterparts theirs was a rougher place,
a less secure business environment. 67 But it was not an impos-
sible one. We even find some examples of officially empowered
merchant organisations in the Roman Empire. One example is the
small association of salt traders in Tebtunis in Egypt, known from
a papyms dated AD47. 68It showsthat themerchants possessed a
monopoly in salt in the area and regulated their mutual relation-
ships. This has been taken as evidence of something approach-
ing strong medieval guilds. 69 That would be wrong, though. The
Egyptian salt merchants had not really acquired a similarly strong,
independent corporate power.Themonopoly existsentirely within
a context of state regulation and, most importantly, state taxation.
It is clearfrom thepapyrus thatthatthemonopoly rights arewholly
dependent onthe merchants administering andpaying the salt tax
due to the state. It was, in other words, totally subordinate to the
fiscal requirements of the tributary system. 70
Political leaders, thus, might both recognise and make use of
the services of traders. But this generally happened in a context
whichvalued urban consumers over the interests of exporters and
the middleman. Illustration is provided by Hadrian's restrictions
on the export of olive oil from Athenian territory. Measures were
taken to ensure that local needs were met before export could
take place. 71 In many respects, as Adam Smith pointed out, such
66 E.g.theletterfromCicero(Fam. 13.75)askingfortheprolongationofsomeprivileges
the grain merchant C. Avianius Flaccus had enjoyed regarding his deHveries of grain
to Rome. For the Mughal Empire, the studies by Das Gupta 1979; Chatterjee 1996;
chapters3 and4; andHasan2004offertelling illustrationsof themore fragilepower
and influence enjoyed by Indian merchants.
67 Bayly 2000 is, to date, one of the best attempts to balance the greater institutional
security of the rising European capitalism with a recognition of the non-passivity of
merchants in the agrarian, aristocratic world.
68 P. Mich. V, 245. 69 VanNijf 1997: 13-14suggestssuchan interpretation.
70 See Weber 1972: 592, 644-645 and 652 for this distinction.
7' SEGXV, 108.SeeingeneralFinley 19853:160-164.ThefamouspassagefromCicero
Off. 1. 150-151, in fact, illustrates the ideological ambivalence in relation to trade:
152
A ROUGH TRADING WORLD
procedures werejust the opposite ofthepolicies that were gradu-
ally introduced during the early modern period in Europe. 72 What
mattered to theGreco-Roman elites, however, wastheavailability
of well-supplied markets in the many cities which made up the
political framework of the Empire and of which they continued
to be the local leaders. Placed under the supervision of the local
aedile or agoranomos, whose task it was to regulate business and
ensure thattraders actedfairly without, asthey sawit, undue spec-
ulation, the markets of the Roman world remained subject to the
"moral economy" of the polis/civitas. 73
Sojourn on the Nile: the Egyptian price series
So far the discussion has been conducted mainly on impression-
istic evidence. Historians and economists examining market inte-
gration, however, are normally used to basing their analysis on
long, continuous series of recorded prices in different locations.
But such statistical analysis is beyondthe means of the historian
for vast swathesofpre-industrialhistory. Recordswere eithernot
collected or have not survived. Only in exceptional cases is it
basicallyintheeyesofthearistocrata sordidactivity,particularlyaspractisedbythose
that buy up goods only to sell them on. Yet if conducted in a'grandmanner,'which
broughtgoodsnototherwiseavailableto market, thenit mightbedeemedworthyof
respect. SeeGiardina 1986for thecomplex ideological websurrounding merchants in
antiquity, thoughhe makestoo muchof a contrastbetweenthedespisedsmall trader
andthehonourablebigmerchant.Thewholesalerwasalsosubjecttosocialprejudice
and suspectedof immoderate greed, speculation andlying, as'is clearfrom'Cicero's
own discussion of the secretive merchant in Off. 3. 50 and"the contempt administered
byPetroniustohisliterarycharacterTrimalchio,thestereotypicalfreedman,vulgarand
greedybeyondmeasure(cf. Sat. 75, 8-76, 9).
72 The theme of the third book oftheWealthof Nations, e.g. Smith 1976, vol. 3: 380,
422. See Bang 2006 for an attempt to relate the problem of Roman'markets" to"the
questions raisedby Smith's understanding of history. The formation of empire should
not automatically be assumedto havecreated a similar strengtheningof commercial
networks;paceErdkamp2005: 196,it gaveriseto a very differentprocessofmarket
formation.
73 SeeAlston 2002: 274-277 and337-339; Garnsey 1988: 257-268; Jones 1940: 215-
219. See Jakab 1997, chapters 2 and 3 for a legal analysis of the functions of the
aedileandagoranomos Thompson 1993,chapter4 contrastedthemoraleconomyof
local communities privilegingthe interestsofconsumers, withthe neweconomyin
eighteenth-century England, which strove to break down barriers between markets and
ensure greater economic integration by siding with the middleman. See Erdkamp 2002
foranattempttoemploy thisconceptinananalysisoffoodriotsintheRomanEmpire.
153
 IMPERIAL BAZAAR
possible to break through the mist ofpre-statisticaltime. The for-
tuitous preservation on clay tablets of astronomical records from
Hellenistic Babylon, for instance, does make it possible to catch a
glimpseofthevicissitudesofcommoditypricesandpre-industrial
trade forlong stretches oftime. 74TheRoman historian canexplore
theextentto whichmarketswereintegratedwithreferencetoprice
material from Egypt, owing its survival to the protection offered
to papyri by the dry conditions of the desert.
No papyri records, however, can match the continuity of the.
Babylonian material. In that respect, the Roman is more fragmen-
tary and episodic. But what it lacks in continuity, the Romano-
Egyptian evidence does to some extent make up for in geo-
graphical variation. The Hellenistic clay tablets, it seems rather
ungracious to observe, are restricted to one city. A significant
feature of pre-industrial trade was the relative unpredictability of
prices. Yet, prices fluctuate in all markets, sometimes violently,
even in quite well-integrated systems. Price instability may be a
symptom oflimited market integration, butisnoabsolute test.Two
other,complementary, criteriaarenormally employedtodetermine
the degree ofcommercial integration: unity ofprices andparallel
developments. Generally, integration ofindividual markets works
as a stabilising influence on prices. The effect of small changes
in local supply anddemand is evened out through exchangewith
other locafities. This, in turn, will cause prices in individual loca-
tions to converge towards the same equilibrium. Near unity of
prices in differentmarkets is a useful indicatorofmarket mtegra-
tion. But there are always costs involved in moving goods from
one location to another. Even in well-integrated markets, prices
are never completely identical; they may vary between two loca-
tions within a span defined by the costs of bringing the two mar-
kets together. If this span is wide, the single markets will remain
largely'autonomous ofeachother. 7-5Ontheotherhand pricesimi-
larity need not necessarily be the result ofcommercial integration.
74 SeeAperghis2004:78-86for an analysis,noting the volatility of the Babylonian
market.
75 Thisisanimportantmodification whichErdkamp2005: 199-200failstoaccommodate
fullyinhisexpositionofthetheoreticalissuesinvolvedinmeasuring marketintegraho^
GibJson'and'S'mout-l995; Stiegler and Sherwin 1985 set out the different criteria for
examiningmarket integration.
154
A ROUGH TRADING WORLD
Coincidence, regulation or some other extra-economic cause may
be equally responsible. A useful additional test of market inte-
gration is whether developments in individual markets follow the
same course. If prices develop in parallel fashion there is a good
chance that it is the outcome of commercial integration. Providing
information on prices in different markets, the Egyptian evidence
makes it possible to explore tentatively all three dimensions of
commercial integration.
The material mainly originates from urban and rural, half-
commercialised, contexts in Middle Egypt. 76 No major market is
represented. This is both a strength and a weakness. It is a strength
because the testimony offered by this material can be taken as
broadly representative of conditions in the vast majority of mar-
kets aroundthe Empire, whichwouldhavebeenno more, in factin
most cases even less, centrally placed in Roman commercial life.
It is a weakness because it tells little about the situation in leading
markets that might be expected to have pioneered a movement
towards greaterintegration, had such a process occurred. Yet, we
are in no positionto choose.The inhabitedpart of the region con-
sisted only of a narrow strip of land following the course of the
Nile andthe BahrYusufcanalrunning parallel to it, andwidening
fora distanceto includetheFayumwithLakeMoeris.Thiswas,in
fact, a physical environment whichoughttohavefavoured integra-
tion. With easy accessto water and only short distances to cover
by land, transport would have been far less of a problem than in
most otherplaces. The relevant evidence spans a period from the
early first to the early fifth century AD. Unfortunately, most prices
in thefirst-to third-centurypapyriareonlymentionedinisolation.
This makes the density of information which can be gained from
these, a few exceptions apart, insufficient to study short-term mar-
ket developments in any detail. The early prices mostly only allow
us to form an impression of the level of prices in the region. 77
76 Rathbone 1991 clearly shows how market trade co-existed withforms ofnatural econ-
omy on the Appianusestate in this area. Farm labourers werenotjust employed on
a free contractual basis. Various ways to tie at least some of them to the estate were
pursued. Wageswould, for instance,oftenhavebeenpaidin kind.
Rathbone1997offersaninterestinganalysisofthismaterial,butoverestimatesitspoten-
tial for studying market integration. The evidence only allows a very broad impression
ofsome long-term trends, cf. Erdkamp2005:204.
i55
 IMPERIAL BAZAAR
In late antiquity, however, the Roman state began gathering
information on prices. Relevant urban professional associations,
the collegia, were asked or required to submit monthly reports to
the authorities about the prices, especially of raw materials, cur-
rent inthemarketplace. A considerable number ofthese individual
declarations survive. 78 These were then compiled and drawn up
in surveys at various levels in the administrative hierarchy, list-
ing recorded prices over a time period for the different names
(districts). Two such price series have been published among the
Oxyrhynchus Papyri, one from the second quarter of the fourth
century and one from the early fifth. 79 Official registration of
commodity prices was intimately linked with the changes intro-
duced to late antique taxation. The government needed informa-
tiononwhichto calculatethevalueoftaxclaimsandconsumption
demands. Forced purchases ofgoods were a regular feature ofthis
system andwere supposed to beremunerated at therates current
inthelocalmarket.Similarly,duesinkindwereoften^converted
intoa lumpcashpaymentaccordingtothegoingmarketprice80
While the broad picture is reasonably clear, some uncertainty
still remains as to the details. It is not known how the collegia
arrived at the prices they submitted to the administration We can
see that most prices did vary over time andaccording to location,
which is an indication that they do bear some relationship to the
situation in the marketplace. But records were "only" submitted
on a monthly basis. Ifprices changed it isunlikely that they did so
only once a month. Presumably the collegia registered the daily
78 E.g. P. Oxy.3624-3626.Forbasicdiscussionoftheseandtheotherlateantiqueprice
rerords."see'the-editor's introductions to P. Oxy. 3628-3636, 3773 and_3624-3626. See
further Lo Cascio 1998andBagnall 2000: 89-9[ decisively contraFikhman 1991-2
(thepricesaremarketprices,notadministrativelyfixedprices).
TO 'P~dxy~. ~T, -l~l3 (dating to 340) and P. Oxv. 3628-3636 (dating probably tothe; early
c.
fifthcentu'ry)arebyfarthe bestspecimens. Thefirstofthesewasprobably producedby
th'elocaTauthorities in Oxyrhynchus, theothercompiled in a provincial officebasedon
material submitted by "thetabularii ofeachcity", asstated in P. Oxy. 3628, lines^1-3.
80 E.g, Justinian,AppendixConstitutionum Dispersarumvii, § 26(inXI
Corpus luris CiviUs
Lo Cascio
mB Novellae, 'ed. Schoell and Kroll), C Theod. VII 4, 10 and 46; 15. 2.
1998removes anyremainingdoubtsastothecharacterofthesedocuments, mey^must
be'understood in the context of late antique taxation practices. Cerati 1975 remains
fundamental on^the importance ofso-called coemptio andadaeratio in thefourth and
fifth centuries. Bang 2007: 31-39 attempts to place the phenomenon in its economic
context.
i56
A ROUGH TRADING WORLD
prices and then presented the authorities with an average or the
predominant price. We must also allow for significantdifferences
in standards here. Not all collegia are likely to have been equally
rigorous, nor will information about all products havebeen equally
important. By far the most elaborate of the preserved surveys, for
instance, documents developments in grain and wine prices quite
closely, whereasthepricesfor less importantproductssuchasmeat
or radish oil are consistently registered as unchanged throughout
the year. 8' The quality of single records is bound to have var-
ied, though it is impossible in individual cases to ascertain with
certainty whether a record is false or not. Even so, none of this
is likely to distort the information so as to render the overall
impression of the market situation unreliable. Generally the image
emergingfrom the dossiers seems to fit quite well with our knowl-
edge of markets in Roman Egypt. Thus the documents do not
fail to register the existence of the government monopoly in the
salt trade familiar from the Pnncipate; they also, as will appear
below, reveal that of the agricultural products wine probably had
the most "developed" market in the province. This is just what
we would expect of the primary cash crop of the estates of the
landowning elite. 82Taken as a whole then andbearing in mind the
poor state of the evidence for prices in most pre-industrial soci-
eties, the Romano-Egyptiangovernmentprice seriesdoprovide an
extremely detailed image of the situation in the markets of Middle
Egypt at two cross-sections in time. Some distortion in the details
is probable, but not enough to invalidate a tentative analysis.
The fifth-century series consists of a centrally drawn-up set of
lists of the price of gold, silver and eight commodities in each of
the nine nomes of the province of Arcadia in Middle Egypt. The
records span one year and the prices are given in four-monthly
Thisismostconvenientlyseenfromthetableonpp. 73-74oftheeditor'sintroduction
to P. Oxy. 3628-3636, summarising the information of the reports. The permanence of
prices of some goods such as meat could very well be a reflection of the existence of
only a very limited market in those goods rather than failure to report developments.
On top of this is the question of tax-inspireci monopolies in Egypt. The salt trade was
a monopoly contracted to private individuals (P. Mich. V, 245), which explains the
uniform rate of salt in the survey.
82 SeeRathbone 1991, chapter 6 on the agricultural production oflargeEgyptian estates;
see further Sarris 2006 on the great Egyptianestates.
157
 IMPERIAL BAZAAR A ROUGH TRADING WORLD

intervals. 83Thismakesthedatainadequateforinvestigatingshort-
term fluctuations, other than in a very broad way. Limiting the
number of recorded prices to three each year for a particular com-
modity inevitably serves to underplay change. This is particularly
so since the prices recorded are likely to represent four-monthly
averages. The calculation of averages automatically irons out dif-
ferences betweenmonths and thus artificially minimises the level Arcadia
of fluctuation occurring in the market. From our present perspec-
tive. however, the set of lists is of particular interest because,
uniquely, they allow us to test for the two other criteria for mea-
suring the degree of market integration: variation in price levels
betweenmarkets andthe degreeof uniformity in developments in
commodity prices across locations. In theory, it should havebeen
possible to achieve a cross-section of prices at three points in time
for all nine names making up the province. But the document, as AphroditODolis
we have it, was never finished. Gaps for three nomes have been
left by the scribe to be filled in later. Wear and tear of the papyrus
has further reduced the available information. The names of two
of the names are lost, as is a considerable number of prices. Of
the eighteen originally recorded wheat prices, only eleven are still Theodosjopolis * .. --/*
Heracleopolis
legible, of wine only twelve. Yet, enough survives to enable a
tentative analysis.
First, let us deal with the variation inprice levels. Comparing the East Mediterranean:
quotations of commodity prices reveals the existence of remark-
ably large differences between the prices current in the individual ^i£ &V-
cities, even within the relatively small area covered by Arcadia,
not more than 100miles in length (see Fig. 3. 2). For instance, the Oxvrhvnchus
difference between the lowest and highest prices simultaneously
listed for meat is 25%, for radish oil, 31%, and 32% for lentils.
Even the value of the gold coin, the solidus, expressed in myr-
iads of denarii, shows some rather surprising differences. Thus
in the period from September to December there was a 5% gap Fig. 3. 2 Roman Egypt and the province of Arcadia
between the price in Cynopolis and neighbouring Oxyrhynchus,
83 P. Oxy. 3628-3633. In working on the papyri I have checked the Berichtigungsliste
and used the images made available in electronic form by the homepage of the Centre
for the Study of"Ancient Documents, University of Oxford (www. csaclox. ac. uk) to
confirmsomeuncertainreadingsofprices in thematerial.In thatconnection,I should
liketoexpress mywarmgratitude to DorothyThompson forspending a most enjoyable
Thursdaymorningwithme,tryingtodecipherthescriptusedbythescribe.
i58 159
 IMPERIAL BAZAAR A ROUGH TRADING WORLD

less than 20 miles apart. This is the sort of behaviourwe would Table 3. i Variation in price levels in Arcadia
expect from a system of imperfectly integrated markets.
This impression is confirmed if we model the market condi- i: Variation span of prices, the difference between highest and lowest
price
tions in greater detail for some of the most important products,
Sept. -Dec. Jan.-Apr. May-Aug.
wheat, barley andwine.They all reveal the existenceof consider- Wheat 88% 67% io8%
able price spans between individual markets (see Table 3. 1). For Barley 30% 78% 33%
wheat, the difference between highest and lowest simultaneously Wine 7% (43%) 15% 27%
recorded price is betweentwo-thirds and, if we can trust the very
low Arsinoite values, a little more than 100% in the year under 2: Commodity prices, real figures (indexed numbers) in the nomes of
Arcadia
examination, though the figure is of course less for some cities.
For instance, the level in the Aphroditopolite name, less than TOO Wheat(den. myr./art.-index) Sept. -Dec. Jan.-Apr. May-Aug.
miles downstream, is 46% higher than the corresponding price Cynopolite 307. 7(ioo) 300(97) 325(106)
Oxyrhynchite 316.7 (i03) -
recorded for the Cynopolite nome in the September-December Uncertain 450(146) 500(162) 500 (162)
interval. In order better to assess this difference we must adjust for Uncertain 450(146)
transport costs. To take advantage ofhigherprices the grain would Arsinoite* 240 (78) 240 (78)
have needed to be taken down the Nile by boat from Cynopolis Aphroditopolite 450 (146) 450(146) -
to reach the Aphroditopolite market. Information pertaining to the Barley (den. myr. /art. -index ) Sept. -Dec. Jan.-Apr. May-Aug.
cost of transport is scanty. But the papyrus evidence we do have Cynopolite 285.7(100) 278.6(98) 300(105)
shows with consistency that the rent of a boat, including some Oxyrhynchite 292. 3 (102) -
Uncertain 270(94) 400(140) 270(94)
basic handlingcharges, is unlikely to havebeen very substantial.
Uncertain
Most graintransport ontheNile seems to havecost less than 3% of Arsinoite 225 (79) 225 (79) 225 (79)
the value of wheat. Cynopolite wheatcould easily have absorbed Aphroditopolite 225 (79)
this extracost. 4
Wine (den. myr. /sex. -index) Sept. -Dec. Jan. -Apr. May-Aug.
Part of the explanation for the very high discrepancy may lie Cynopolite 20. 6 (l00) 20. 6(100) 22.5(109)
in that pronounced enemy of market integration, the levying of Oxyrhynchite 22.5 (109)
local customs (see further Chapter 4). We know, though the evi- Uncertain 21. 4(104) 23. 6(ii5) 28. 6(139)
dence is heavily weighted towards the high empire, that it was Uncertain
Arsinoite 20 (97) 23. 8 (ii6) 27. 5(133)
normal for customs dues to be paid on exports from and imports
Aphroditopolite ? 28.6 ? 28.6(139)
to each nome. Whether this also included paying for goods in
transit on the Nile each time a boat crossed a name boundary is
* The grain prices in the Arsinoite nome may be suspiciously low since
they also fail to register any developments during the year. But this is
84 See Adams 2007 for a recent analysis of Egyptian transport. The price evidence is a matter for speculation. Therefore these data have not been excluded
collected by Johnston 1936 andDrexhage 1991. P. Oxy. 522 (second century AD) gives from the calculations. Exclusionwould, in any case, not alter the basic
a pure freight rate of 21 drachmae (dr. ) per 100artabai for Oxyrhynchus to Alexandria. picture.Thevariationspaningrainpriceswouldbelessened,especially
Setting wheat at a reasonable 8 dr. per artaba this comes out at 2. 5%. The boat, dealt for wheat,but it wouldstill behigh.
with in the papyrus, carried 3,400 artabaimaking the freight rate 712 dr. To this were
addedvarious handling and loading charges,bringingthe total bill to 762 dr. 3 obols
(ob. ) or 2. 8%forwheatat 8 dr./artaba.Otherpapyri statefreightrates inkind,e. g. BGU
Ill, 802,col. 22 givingpricesfor boattransportfrom theArsinoitenometo Alexandria
ofthree differentagriculturalproducts. All come out at lessthan I % ofthecargo.
i6o i6i
 IMPERIAL BAZAAR A ROUGH TRADING WORLD

unclear. The craftiness of custom officials is legendary and, to

Priceof wheatin fifth-century provinceof Arcadia
judge from comparative examples, it would not be at all surprising
if transit boats were charged too, possibly at a lower rate. The In myriads of denarii per artaba In indexed numbers
Cynopolite name is separated from the Aphroditopolite by the nome
Sept.-Dec. Jan.-Apr. May-Aug. Interval 1 2 3
Hermopolite. Customs would then at a minimum be exacted on Cynopolite 308 300 325 Cynopolite 100 97 106
Uncertain 450 500 500
departure and on entering Aphroditopolis. The going rate seems Uncertain 146 162 162
Arsinoite 240 240 Arsinoite
to have been around 3% each time. 85 To the approximately 3% Aphroditopolite 450 450
78 78
Aphroditopolite 146 146
needed for transport, this might have added an artificial barrier of
at least 6%, or more if dues also had to be defrayed for transit
Histogram of indexed wheat prices
through the Hermopolite nome. Neither cost would in this case
have been prohibitive. But the need, willy-nilly, to pay perhaps
up to 10% of the value of the grain in the process of bringing it
to market may have discouraged the establishment of a routine
iiCynopolite
commercial infrastructure, able to take advantageeven of smaller
H Uncertain
price differentials. DArsinoite
The impression of relatively low andfragile integration between . Aphroditopolite
the individual markets in the province of Arcadia receives further
confirmation when we proceed to compare the development in
grain prices over the recorded year. Unfortunately, the damage to
thepapyrimeansthatnotall priceinformationhasbeenpreserved.
Wecanonly follow theprice during all three four-monthly intervals
for Cynopolis and a city whose name does not survive (Fig. 3. 3).
This and the fact that there are only three prices recorded during Development of wheat prices
the year, mean that the result must remain very tentative. From 200
the graph it appears that price developments do not conform to the
same pattern. While prices stay essentially the same in Cynopolis 150 4
in the second interval, the price increases 11% (index price from
I
100 -I
.

Cynopolite
146to 162) in the unidentifiedcity. In the third interval, the pattern Uncertain
is turned around. NowCynopolis experiencesan increaseof some
8%, whereas the price remains unchanged at its high level in the
I 504

anonymous city. The situation for barley puts this pattern into
0
sharper relief (Fig. 3. 4). The uncertain city experiences a sharp 1

rise from the first to the second interval and Cynopolis remains .

Cynopolite 100 97 106

stable. By the third interval Cynopolis shows a slight increase in - Uncertain 146 162 162
4-monthly intervals
85 See Sijpesteijn 1987: 19-25 and de Laet 1949, 324-329 for customs rates in Egypt.
Sijpesteijn 1987: 6-7 is not justified in taking the virtual disappearance of evidence
pertaining to customs from the third century onwards as a sign that the exaction of tolls
Fig. 3. 3 Price of wheat in fifth-century province ofArcadia
had been discontinued.

162
i63
 IMPERIAL BAZAAR A ROUGH TRADING WORLD

Price of barley in fifth-century province of Arcadia the price ofbarley, whereas thecity ofuncertain identity now has a
In myriads of denarii per artaba In indexed numbers steep fall inthepriceofbarley. Thepricesareveryfarfrommoving
in unison and market developments appear to vary considerably
3
between the cities.
name Sept. -Dec. Jan. -Apr. May-Aug. Interval 1 2
Cynopolite 286 279 300 Cynopolite 100 98 105 Perhaps, a little surprisingly the harvest cycle fails to manifest
Uncertain 270 400 270 Uncertain 94 140 94
Arsinoite 225 225 225 Arsinoite 79 79 79 itselfclearly. Themain grainharvest in Egypt seems normally to
have taken place in May. Prices could therefore be expected to
rise in the second interval of the year as stocks began to wear
Histogram of indexed barley prices
thinner, and drop again in the third interval after the new crop.
Only the unidentified city conforms more or less to this pattern.
But its wheatprice remains high for the May-August interval. A
possible explanation would be that the harvest came late in this
1 Cynopolite
M Uncertain
particular year. The delayed rise in the Cynopolite grain prices
could then be a reflection of this and perhaps also indication of
DArsinoite
a bad or disappointing crop. 86 Another likely explanation might
be distortion caused by the method of registration. However, ' if
market integration was fragile, we should expect local deviations
from the "standard" pattern. If it had been possible to test for a
muchgreaternumberofcitiesandperhapsovera longertimeframe,
and the average pattern still failed to conform to the rhythm ofthe
Development of barley prices agricultural year, it would be a greater cause for concern. In this
case, however, the number of examples is so small that individual
150 -T
irregularities are only to be expected and may, in fact, be a result
of market imbalances.
I ioo 4 -^-Cynopolite
3 -*--Uncertain
Finally, let us consider wine. In contrast to the grain prices, a
-A-Arsinoite
very clear harvest cycle pattern comes through in the evidence;
I 50 -)-
prices are low in the post-harvest interval (September-December)
0
and rise steadily towards harvest time in August, reaching cul-
1 mmation in the the third interval. This requires introducing extra
-»-Cynopolite 100 98 105 subtlety into the analysis. Shared patterns of price develo'pment
-a-Uncertain 94 140 94 cannot, in and of themselves, be taken automatically as evidence
-ik-Arsinoite 79 79 79 of close integration. They may not be due to a close coordination
4-monthly intervals of supply and demand across marketplaces. Equally important
may be exogenous influences common to all the markets, such
Fig. 3.4 Price of barley in fifth-century province ofArcaclia as the arrival of a plague, political collapse or shared weather

Rathbone199^1:228-231documentstheprevalenceofa harvestpatternlaterthanthe
norm for the Fayum.

164 i65
 IMPERIAL BAZAAR A ROUGH TRADING WORLD

conditions. This last factor, by producing similar harvest results Price of wine in fifth-century province of Arcadia
across the region, is undoubtedly part of the explanation for the In myriads of denarii per sextarius In indexed numbers
shared pattern in the wine markets. However, wine was the pri-
mary cash crop of Egyptian agriculture, and we should probably nome Sept.-Dec. Jan.-Apr. May-Aug. Interval 1 2 3
bepreparedto accept some integrationof markets. Cynopolite 20. 6 20. 6 22. 5 Cynopolite 100 100 109
Uncertain 21. 4 23. 6 28. 6 Uncertain 104 115 139
Thus, even though their courses are not absolutely identical, Arsinoite 20 23. 8 27. 5 Arsinoite 97 116 133
with the price leader changing twice during the year, the curves
for the unidentified city and Arsinoe do appear to show some Histogram of indexed wine prices
resemblance.87 On the other hand, the fragile and limited nature
of the integration in the wine trade is clearly brought out by the
overall development in the market. At the beginning of our period
the prices of the unidentified city, Arsinoe and Cynopolis are very
close together within a range of 7% from bottom to top, while at
the end of the year that gap has increased to 28%, and Cynopolis . Uncertain
has seen a significantly different development than the other two HArsinoite
markets (Fig. 3. 5). Incidentally it is worth observing that local
customs duties might have been an important barrier to integration
in the wine markets. All in all, this leaves an image of a set
of markets where integration was limited, links were fragile and
vulnerable, and coordination between the individual markets fairly
weak. As a result conditions varied considerably from city to city,
even within this very limited territory. Development of wine prices
Moving on, a fourth-century papyrus lists information on the 160
monthly, presumably average, prices for gold, silver, wheat,
140
barley, lentils, vegetable seed and wine over a fifteen-month
120
period (see Table 3. 2). 88 As was to be expected we find a pattern 5
for the agricultural goods heavily influenced by the seasons with a 100
E .

Cynopolite
3
pronounced tendency for prices to peakjust before harvest and to 80 Uncertain
slump with a little delay after. But the agricultural seasonal cycle I 60 -*-Arsinoite
cannot account for all developments. Again some irregularities, 40
now less suqwising to us, emerge. Other factors clearly also had 20
a strong influencein the shortterm asthe monthly averagesshow. 0
87 One reason for their close resemblance could be that the unidentified city was located 1

close to Arsinoe, e. g. if it was identical with Herakleopolis at 26 km's distance, which Cynopolite 100 100 109
might well be the case.
Uncertain 104 115 139
88 P. Oxy. 3773. The editor assumes that the prices are real market prices. However,
if government officials could calculate four-monthly averages, I have no doubt that Arsinoite 97 116 133
collegia would be able to produce average monthly prices too. In view of the violent
fluctuations of some of the prices, it would be highly unlikely that prices within each
month were absolutely fixed. Fig. 3.5 Priceofwinein fifth-century province ofArcadia
166 l67
 A ROUGH TRADING WORLD

Q Thepriceofbarleyisrevealing. Itsetsoffwiththeexpectedrise
^f:
£:'
IS asstocks wearthinner during winter. Then inMecheir it suddenly
drops by 20%, only to rise to its former level in the followins
^ s .- §'
month. The harvest sees a slump but only for two months, foL-
Ill loweda little surprisingly by a steepincreaseof33%. Thescript
Ni for wheat is slightly different. It shows an 11%'increase from
^ ^Sl Phaopi until Pachon, then drops in Pauni to the previous autumn
u t-~ S
-t M
^ &
level as the crop is harvested. But then in That a fall of another
^ ^I i ll% follows rather unexpectedly. For lentils, a similar pattern of
&<
^ Is some erratically fluctuating prices appears. Here, as for barley,
0
s
il
Q ~a
there is a surprising drop in Mecheir (12%) and the fall in prices
^ ^ c^ t~-
afterharvestisveryquickly offsetbya largeincreaseinPauni.Did
i the crop fail? Or was there some other shock, unknown to us? Be
i
-S; that as it may, the situation seems to have been unstable - as the
.

^ further 17% increase in Thoth and its immediate nullification in

Phaophi indicate Something similar couldbesaidaboutvegetable
3 I ^ seed, where weobserve anincrease ofsome 60% during the time
0
§i wecan follow the price. From these observations it should beclear
e r<~) .-0
that seasonal and other changes combined to make the markets
t c". '2
quite volatile during the year. Further, developments in individ-
i IS
i £ .c
ualgoods showfairly divergent patterns, anothersignof weakly
integrated markets. That becomes even clearer when we include
^ &1 5
~Q gold and silver in our observations. The document stems from
I^
in in >n 0
ttieperiodofconsiderable inflationaffectingtheimperial coinage.
The timeframe is not long enough for this to have a visible effect
.1 on the prices registered for the agricultural goods. But with gold
^
and silver it is another matter. Both seem to rise in Athyr. Then
a,
^
r^ m 1st
§5-5
silver sees an additional increase in Mecheir. In Pachon both silver
~§ s^ .
: II andgoldrise thoughnotbythesamepercentage. Verybelatedly
'£. & f-
QS| goldseemstofollow suit(fourmonths later, infact), whereassilver
climbs to itshighestlevelhitherto inThoth. Insum,thefourth- and
a ^ .s S
fifth-century government pricelistsconveyanimpressionofvery
M

m
-D' . 5 JS <?
opaque, weaklyintegrated markets withrelatively frail stabilising
u
'» 3 S S Ill mechanisms.
I i a a. s-^s-s Consisting ofmonthly andevenfour-monthly priceaverages, the
a
a !a a fourth- and early fifth-century lists must invariably have masked
f
"3
w -2 ^
u JU
some short-term fluctuations. Some ofthis wecanmake upforby
complementing ourimageofmarketsinMiddleEgyptbydrawing
169

IMPERIAL BAZAAR
in some of the earlier, but more disparate information on prices
w'e'D'os'sess from'the area. 89 Effectively we have only a handful
Jdocuments offering sufficiently detailedevidence onwhichto
create short-term price series. Several oftheserelate to^thepnce
o^win^This'cames problems ofits own. Often it is^not possible
to ascertain whether a variation in price reflects different quali-
ties of wine or changing prices. 90 A couple of other series come
do
from'internafe state accounts. While there is no doubt that they
reflect market prices, asarguedbyRathbone, overthelonger term,
itis'farfrom dearthat they were adjusted according to short-term
developments in themarket. Thus one showsaseasonal
surPnsm^stat)u-
"that 'does not even register the expectable changes.
There^'for instance, nodrop inthemonths aftertheharvest. To
eefrom thefigures, the level ofthe accounting price seems to
JhavfeeYeenadjustedannuallyinJuneandmoreorlessmaintained
for the nexttwelve months.9'
This leaves us with the information that canbe s<
theTebtums^p/^on, orrecordoffice,forSeptember andOcto-
ber AD> 45.92 Here we find a number of transactions for the
M H.-J.Drexhage 1988and 1991 andRathbone ,997havegatheredAematerial
: Sl;SSEji^iijis=
S;i^Siii^|^?S£S
^e^^^hb^T997:^:Thisproblem, ^c, d^^^J>^^
^e^^^''hcP^^^tte^inedataProiectafairlycohercntJ'TClge'Th£tp^
^S^^s^'^uffi^n^sigm^amtorequireexptanauonj. ^refl^
^S^^^^h?:^'Sm 'ES^
S^iZ^'and ^^lik£>ytavt, ten, b;^n^^'^^: The
d^'thereforewere theresultofa muchmorehomogeneous co".ectio"Pro(^dure_
9. ^^^^d^Rad^ne 1997:229-230).TheH^ation^mN^n^
91^S^^^'^^n^od^Th^^o^^^a^
K^'^^no'infom^ion forJunebutinAugustittasns to^d^tA;
^^:^^nc<. ta^ln0ctob^^^^Tghtl^b^^^^
H^Si^ijij^^cSS
sSijii'i'iHiiS^ss
^=^iiiss=s;s:
the'thirTcentury,wouldoftenhavebeenadministrativel^fixe^ _^_^^^
^ Sl ^iT^"^ gramp;i(:ei^M"/LI^2£^^^
^^r^e^emo^^^emenriy'listedin^uncan-Jone^ 1990^becauseheliststhe
gra"mprrices"according todate.SeealsoRathbone 1997:217and224.
170
A ROUGH TRADING WORLD
procurement of small quantities of wheat and wine. Taking wheat
first, on 7 September in AD 45 the price is listed at 4.4 drachmae
(dr. ) per artaba. A fortnight later it has risen by 30% to 5. 7 dr.
Three days later it is 8 dr. or an extra 40% higher. 93 Two days
later it drops to 7. 3 dr., that is by 9%. In October it goes up in
two instalments to 8 dr. again. It is clear that even in the very
short term, prices fluctuated quite violently at times. This conclu-
sion is further supported by the data on wine from the grapheion.
In the same months, the prices for wine fluctuate within a band
of 2. 50 dr. and 4. 58 dr. per keramion or by up to 80% in round
figures.94 This image of fragile market integration would appear
to clash with Rathbone's analysis, partly based on the same data.
By a comprehensive statistical analysis of the fairly few attested
prices for wheat, wine and donkeys from Middle Egypt in the first
three centuries AD he concluded that markets seemed to be well
integrated in the region. His conclusion was based on the identifi-
cation of a common long-term pattern with two periods of relative
price stability within relatively broadprice bands from c. AD 40
to 170 and again from 190 to 270. 95
First, the price bands are very broad. That for wheat in the first
period lies in an interval between 5 or 6 and 12 dr. per artaba.
This leaves plenty of room for local intra-annual variations. If we
look at the very small amount of data that constitutes this band
we find that it is essentially basedon ten prices from the Arsinoite
93 One possible explanation for this extraordinarily high leap could be a very poor Nile
inundation. Prices will therefore have soared in response to the expectation of a bad
harvest in the coming year, Rathbone 1997: 193. However, that remains only one
possible hypothesis. If we are to judge from slightly later material, the start-out price of
4. 4 dr. per artaba was very low. We could therefore also just be seeing a return to more
normal prices.
94 Pace Ralhbone 1997: 196-197, who concludes that changes over the very short term
were minimal. To support this conclusion he calculates the variation from the median
(the middle point between the two extreme points in the price range). Over the short
term that is going to produce a false sense of continuity. For instance, a price rise of
100% is only goingto appearasa variationof33% aroundthemedian(example: price
goes from 5 to 10. Variation around the median will be only 2. 5 around 7. 5 or 33%)
and a rise of 50% will come out as a fluctuation of 20%. The merchant, however, is
not going to pay the median price. His profit is going to depend on the variation in
real prices. Cf. Steensgaard >973a: 42-59 for a similar approach to the calculation of
short-term fluctuations in pre-modern markets. Working on early modern trade he even
considers as violent much smaller variations than our evidence shows.
95 See Rathbone 1997: 210-216fora summary ofhisresults. H. -J.Drexhage 1988 is much
closer to the present author's analysis.
l7l
 IMPERIAL BAZAAR
nome andsevenfrom the Hermopolite. Takenseparately, thedata
produce an average of 7. 36 dr. for the Arsinoite and 9. 55 dr. or
30% higherfor the Hermopolite. Thisdifference may becoinci-
dental. Ten and seven figures respectively from a period of over
loo years canhardly beconsidered adequate evidence to allow
conclusions regarding differences in the price levels between the
two names. But they'do remind us that we cannot exclude local
differences within thevery broadtrends located by Rathbone. In
tact.forourfifth-century dossierbasically all thewheatpricesfall
within a similarly wide span, and yet considerable fragmentation
characterised the market; The broad comparability of early and
lateprices, incidentally, corresponds quite well withrecenttrendsth
inlate antique scholarship stressing the continued
economic
heal
andvitality ofmarkettradeinthefourthandfifthcenturies. 90The
evidence of short-term fluctuation and the late antique records,
therefore, both suggest that the parameters, never made explicit,
of Rathbone's analysis are faulty. The Egyptian prices from the
Principatearefullycompatiblewitha scenariowheremarketinte-
gration remained fragile.
Second, it is important to recognise that this is not a question
of either total disintegration with completely isolated communi-
ties or fully correlated markets moving in absolute unison; we are
trying to look atdifferences between a group ofhighly complex
pre-industrial societies with forces working on several levelsto
tie localities intowidernetworks. Theactivities ofthestatewould
have been one significant factor in creating some broad overarch-
ingpatterns, aswasarguedinthelastchapter.Ona micro^level
state-defined rates for compulsory purchase or the conversion of
graintaxes into cashpayments may havehelped stabiliseprices
inthelongterm withinthebroadbandsidentified byRathbone. 97
Compulsory purchases, of course, would often have served to
96 Bagnall 1993,Banaji2001,Sams2006andZiche2006.ITheyall'"sistand
onAein
continued
3v'o7e7 cJon"om^liteYn ate
much of the Empire during antiquity, particular
Egypt'The7efore"it'is"not unreasonable to usethismaterial, atleastasanindicator .
conditions in the preceding centuries.
97 SeeRathbone~i997: I96-W t'orthissuggestion. Statepurchases andc°"v^onrate^
^n onTv'hav'e be7en' very'broiid 'stabilising'factors. According to Rathbone's P"ceband^,
rtie'spanofvariationsbecomesnarrowerinthethirdcenturywhenevidenceofconversion
stops for a period.
172
A ROUGH TRADING WORLD
increase uncertainty in the short run, as the many surviving com-
plaints clearly testify; often people wouldhavesuffered losses and
been forced to sell below the current market rate. Individual com-
munities did not exist in perfect economic isolation either. Trade
clearly didtakeplace. Inthe long run this must haveworkedto cre-
ate some broadly coherent features. Even if prices in one Egyptian
city would normally have differed from those of its neighbours,
we would expect the variations to remain within limits. If they
had consistently been of an entirely different order, landowners
in the surrounding area would surely have moved in to benefit
and thus gradually re-established a connection, however broad,
between individual prices in the region.98 The identification of
some common long-term price trends, therefore, does not militate
against the case for a relatively low level of market integration.
They are only to be expected. " Rathbone, for instance, convin-
cingly suggests that the caesura between the two identifiable price
bands was caused by the dismption that followed the arrival of
the Antonine Plague in Egypt. '00 A similar leap in prices can be
observed across large sections ofEurope after the shock caused by
the Black Death. That was long before the arrival of more closely
integratedmarkets. Ratherthantheresult of a closeintegrationof
supply and demand, the jump in prices after the Antonine Plague
wascausedby anexogenousshocksharedby the wholeprovince.
The limited institutional coherence of the tributary empire
So, the question is whetherthe Roman world ever made the move
beyond weakly integrated, volatile markets with some degree of
long-term coherence to more closely correlated and stable mar-
kets. The Egyptian evidence surveyed, so far, suggests a negative
answer. But this answer could admittedly be better supported.
The evidence is disparate and spans approximately 400 years.
98 Steensgaardi973a:58 arguesthatshort-temiweakintegrationofmarketsdoesnot
exclude the creation of broader conjunctures over the long-term.
99 Grover 1966 provides an analysis of that sort of low-level integration in the Mughal
Empire.
'°° seeScheidel2002forananalysisoftheeffectsoftheAntoninePlagueinEgyptseen
in the light of the Black Death.
173
 T
IMPERIAL BAZAAR
A ROUGH TRADING WORLD
Significantchanges took place within that period. On top of this,
the information does not originate from some of the more impor-
wouldbearistocratic households. Theyoftencontrolled enormous
tant markets in cross-regional trade, where the pulse beat quicker
amounts ofagricultural produce forsale,atleastsufficient tomake
and closer integration was more likely to make its first appear-
themimportant players ona localorregional scale7However"l
ance. However, we do have a few scraps of information issuing
did"<,)tdTend
rested their
°"theirabilitytotiemclrketstogetherThei!:'powe'I-
on position in the political hierarchy and in the con trol
from people in touch with more central markets. One of those is
provided by a letter sent from Coptos in the early second century.
oftheagricultural surplus. Indeed, often they seemrather to have
The city held a central position in the long-distancetrade between
exploited thisposition tobenefitfromenhancingtheendemic mar^
Egypt and India, and developed into an important interregional
imperfections; I have already mentioned the resentment ofthe
market in the province. 101 The sender had been commissioned by
wealthyd.assesinAntiochtothevariousimperialattemptUo'all'^
a certain Apollonius to procure a number of goods there to the
viatea particular foodcrisis.Libaniusgavewordstothe'sentiment
underlying theresistance. In a letter toa high-ranking offic'iai"t'het
substantial value of a little less than t , 000 dr. The letter, there-
fore, offers an interesting glimpse of conditions in a fairly impor-
count of the Orient, Rufinus, he only grudgingly admitted'that
tant commercial town and may help us to fill in the blanks. The
^interference in a food crisis hadbeen necessary'and that Ae
writer tells of his careful efforts, mostly crowned with success,
interference with the "autonomy of the marketplace7'"hacTbee'n
to secure a good bargain for his correspondent. At the end of the
tlwlesserevil. '^often landowners wereunwiTlingto matee;en
business report, he declares his willingness to undertake further
thatconcessiorLSomegroupstooktheirresistanceotointe7vention
commissions on behalfof Apollonius andthen addsa bit ofexpert
^thewayto theemperor's court: a ruling ofMarcusA'urelius'
information: "the price of bullion is now 326 dr., for as you know,
andVerusfelt obliged toemphasise that"decurions'shouldnotbe
the prices at Coptos change from day to day". " Again, it is the
'. to.provide. graintotheirdtizensatlesstha"theprevailing
impression of endemic and erratic local price fluctuations that
market. price"-, '05Ingeneraltheelitedldnotlookfavourably (
seems to receive confirmation. This image also conforms to the
attempt to reduce its capacity to exploit its economic resources'.
^aramg andspeculative manipulation ofthemarket bythenower-
general expectations of the Roman elite. According to the second-
century jurist Gaius, prices of many important commodities such
fill, notleast in thenecessities oflife duringtimes of"cr-^nd
in the ancient
as wine, oil, grain andeven credit, were likely to showstrong local r,Tl n!lu.l!commonplace sources.
'06
It speaks vol-
ume^thatamongthe/elativelyfewdirectionsabout'ev'e'rydaylife
,

or regional variations. 3 ,

We can go further by contemplating whether it is possible to

in the Lex Irnitana (a Spanish municipal law dating to'th"e late
pinpoint any institutions with the capacity and sufficient power
to be able to establish the long-term control of markets necessary l^aniusEp.1379,2 ("TTIVayopavauTovo. ov")withGarnseyandWhittaker1998:
to make them more closely interdependent. An obvious candidate "s ^t, 30^(Marcianus)LFOr thlsanda fewsimilarPassages, seemostrecently Lo
=S^?SSIi^i=S
"" Plin. Nat. V, 60 on Coptos. For a recent collection of papers, see Autour de Coptos.
102 Translation, Johnston 277 of P. Giss. 47 (= W. Chr. 326). The text, however, must
be read with the corrections to line 29 published in the first Berichtigungsliste by
S^i=sii^iiSS
Preisigke from 1922.
"'3 Dig. 13.4. 3. Similar is Pliny Nat. XXXIII, 57. One might also adduce more literary
forms of evidence. To satirists. novelists and philosophers alike trade remained an ^i^m hisforty-sixthspeechharpingonabou^seif^on^t^ Sn^ ^)^
uncertain activity, a game of chance first and foremost. See Erdkamp 2005, chapter 4
for a broadly similar assessment of integration and fragmentation in the Roman grain E'g', phil os, tratus v't:. APOtA:I5 (tre;lted by Gamsey '988: 76). The Elder Cato i
trade. 3^ ^ I^^, ^e^riS[S;
ablel t^^SS
sufficient storage capacity be
tohoa^^u^al^^^<hS
174
175
 IMPERIAL BAZAAR
first century), we find a sombre prohibition against any kind of
hoarding. 107Thepoliticalelitenormally possessedsufficientsocial
power to be ableto benefit from the volatile market environment
and the endemic, sometimes wilfully created, imbalances in sup-
plies. It had no incentive to try to change things by investing
in greatermarketpredictability instead. 108At times, exploitative,
short-term speculation is even presented asa convenient solution
for aristocrats in need of more cash.I09 This was an attitude the
Roman elite shared with its later Mughal counterparts. 110
Moving a step further up the ladder we find another possi-
ble candidate to-exercise a stabilising and integrating influence
on markets, the different segments of the imperial state. Faced
with the danger of social upheaval, cities, provincial governors
andtheemperor occasionally all hada strong interest in checking
the worst abuses. '" But such intervention, as pointed out above,
happened with a view to local conditions, not market integration.
At anyrate, most initiatives werediscontinuous, implemented on
an ad hoc basis and only during times of severe crisis to ensure
social peace. "2 The bigexception, ofcourse, wasthefood supply
of the imperial capital Rome, perhaps followed by some precau-
tionary measures of more limited help in other large cities of the
. °7 LexIrnitana,section75. (ThoughnotthemostrecentLatinversion,Gonzales1986
offers aneasily accessible edition andwith a parallel Englishtranslation.)
.
°8 Erdkamp"2005:155-74makestheimportant'pointthatRomanlandownersoperated
witha'tewca'rry-over of grain stocks'from year to year. Whenprices were volatile
it madebettereconomicsenseto profit on highpre-harvestpricesthanon long-term
storage. Biglandownersthereforewerepartofa regimewhichreproducedmarket
imperfectionsandfragmentation,paceHobenreich1997:268-282
109 Suet.Vesp. 16 i. Whittaker 1985suggeststhatthereisa stronglinkbetweenmember-
ship'oHheimperialeliteandimportstoRome.Presumablysenatorsexploitedtheir
powerto gainadvantagesintheRomanmarket.
''" A'li"1997"chapter6.forexamples.Inthatchapterhealsotreatsanimperialfannan
(edict/clenouncing the manipulation ofthemarket by Mughal nobles.
.
" Apart'from theLe^/mttem(/(GonzaIes 1986)(themunicipallevel) onecouldp^
evidencefrompapyriandinscriptionsshowingRomangovernorsorderingthat^tocks
be'made available for sale in the market, not exceeding a specified price^e. g. Lucius
AntistiusRusticus'measuresfora grainshonagein PisidianAntioch(AE 1925, [2&,
trans. Sherk 1988, no. 107). ^ _^ , ^ ^ _, ^
"2 TheDhenomenonhasreceivedfull treatment in Gamsey 1988.withslightlydifferei
"lasis in Erdkamp 2005. chapter 6. Seefurther Grenier 1997 tbr some perceptive
commentsonhowadhocattemptsto curbthemarketwouldhaveworkedto increase
commercial unpredictability ratherthan improve thefunctioning ofmarkets.
176
A ROUGH TRADING WORLD
Empire, such as Alexandria and Syrian Antioch. 113 From taxation
the imperial state controlled enormous amounts of grain, sufficient
to make it by far the biggest player in the field. This allowed it to
organisethe grain supply of the imperial metropolis into a much
more settled and stable pattern than was achieved in republican
times. "4 This did undoubtedlycreatea more predictableenviron-
ment for the Roman consumer and secondary private suppliers.
But the result wasto a largeextentbroughtaboutby crowdingout
the market. A very large part of the grain was simply handedout
free to the citizen population as a privilege and what was left in
the imperial granaries would have been sold, probably at a fairly
low market rate. "5
"3 Sharp 1998: 127-140 for some observations on a scheme meant to benefit the food
supply of Alexandria.
114 The grain supply of Rome has received extensive treatment. See Pavis d'Escurac
1976;Rickman i98oa; Garnsey 1988; Herz 1988;Sirk.s 1991; Hobenreich 1997; and
Ercfkamp 2005, chapter 5.
' '5 The Roman state seems to have handed out grain free ofcharge to 150, 000 to 200, 000
people at anannual rateof60 moclii per person. At a generous annual consumption rate
of 175 kg wheatperpersonthiswouldhavesufficedto feedwithgraina populationof
350, ooo to 450,000 in very round numbers. In other words, via the grain dole alone
the Roman state probably satisfied the grain needs of not less than one-third of the
populationofthecapitalandprobablycloserto one-half.Ontopofthismustbeadded
state saleson the"free"marketin thecapital.Regrettablyweareleft very muchin the
dark by our sources. We should not imagine a static picture. The proportion of state
sales in the market is likely to have increased asthe state apparatus in the capital grew
during the Principate. I suspect that during the Principate, the state would relatively
quickly have become a very important supplier of the "free" market as well. From
taxes, the state commanded far more grain than it needed for both the army and the
free handouts. Rome would have been a convenient place to sell some of this. It is not
necessary to imagine the creation of a strong centralised commercial administration
in Rome. Documents in the banker's archive found outside Pompeii, the so-called
Murecine tablets, show us that consignments of Alexandrian wheat were released on
themarketvia small, individualprivatemiddlemen(TPSulp.45-46and51-52,dating
to the reign of Caligula, mention the giving of 13, 000 modii of Alexanclrian wheat,
stored in a Puteolean granary, as security. Dig. 39. 4. [0. 8 confirms that it was normal
procedure for the state to avail itselfofprivate merchantsto sell the produce which it
controlled). A letter of the emperor Haclrian (Oliver 1989, no. 187) refer to the state
shipments of Alexandriangrain as intended for the market ("Trpos TT|V ayopav") of
the capital. This may suggest that state sales were routine, which is why Nero hoped
to gainon highgrainprices in Rome (Suet. Nero45. 1). Finally, reports ofan episode
of speculative hoarding of state grain during Commodus' reign seem to imply that
state grain played or had come to play an important role in the market (Herodian I,
i2, 4 and Dio Cassius 73, [3, l). Rickman l98oa seriously underestimated the role of
the state in the grain supply because he assumed unrealistically high per capita grain
consumption. See Gamsey 1988, 231-239 for a general, albeit on balance probably
still too conservative, discussionof the role of the state in the Roman grain market;
177
 T
IMPERIAL BAZAAR
The emperor did not run a commercial establishment Moye-
ment of grain to the capital happened independently of market
considerations. It wasa question ofstate. Theemperor couldnot,
if he valued his own security, let increased demand in other areas
oftheempire havea decisive effect onthesupplies broughtto^the
Roman market. As Hadrian wrote ina letter tothecouncil ofEph-
esos,~onTywhentheneedsoftheimperialcapitalhad[beeniserved
wouldothercommunities beallowedtobuystategrain. "6 A bad
harvest ina grainsurplus-producing province suchasEgypt, caus-
ingpricesto risein Alexandria, wouldnotbeallowedto have^a
strong effectonthemarket inthecapital. Iftheemperor hadgrain
to spare,hemightopenthestategranaries[nMey:mdnamevm
shipbackstocks from Rome andsell it atreducedrates. Butthis
would only be doneto avert disasterandthedecisionwastaken
withoutregardtocommercial considerations. "7Thusitisunlikely
that'there ever arose a very close link between the situation in the
Alexandrianand Roman grain markets.
"In years of abundance, on the other hand, the imperial state
could dispose of far more Egyptian grain than it needed for the
cap'itafand the army. Some"of this would have been exported
from Alexandria to locations in the eastern Mediterranean expe-
riencing a temporary or permanent graindeficit.This would, to
some extent, have exercised a stabilising influence on the Levan-^
tin'egrain "market". However, asthe above-mentioned letter of
HadrTan to Ephesos goes on to show, being admitted to the mar-
ket in Alexandria was a matter of acquiring a political privilege
from the emperor or more often the prefects of either Egypt or
o'fthe grain supply. Much else other than commercial considers
tions was bound to enter into the decision whether access should
begrantedornot. Patronageconnections andpolitical influence
Erdkamp 2005, chapter 5 f°r a more generous estimate. As the^bi^ ^ww[
m'Ave"e'mpu:e,"the"imperial statewasby implication a majorsellerinthemarket, as
recently argued by Lo Cascio 2006.
116 Oliver 1989, no. 187. ^ _ .. ^ . ^ _ ^_:.
. -7 SeeP\m. 7an. 30-32onthere-shipment ofgraintoAlexandriatoaverta famme_
fact'that'thi's'occurrencereceived'extensivetreatmentin a panegyricto th^empj-ror
E^es"to"stow"howunusuaritmust have been. The same impression i^convey^edl to
^it u7ind^n'ant repo'rt'ofGermanicus' opening of the granaries in Alexandria
alleviatehighprices, Ann. II, 59-
178
A ROUGH TRADING WORLD
would not have been the least important elements in this selection
process. Josephus relates how Herod the Great, during a famine
whichaffected large parts of the Levant, was able to secure a con-
signment of Egyptian wheat by exploiting his connections with
the prefect of Egypt, C. Petronius. For he, "thoughmany persons
had come to him becauseof the same needs, was a friend of Herod
and wished to rescue his subjects, and so he gave them priority
in the export of grain". "8 Access to the market in surplus grain
was, in other words, far from open or equal. It is not impossible
to imagine that preferential treatment will often have increased
market imbalances in the region, ensuring supplies where they
were not urgently needed. Therefore it seems implausible to look
to the Roman state, though it might quite possibly have created
more stable conditions in some markets, for the creation of a set of
tightly integrated markets between the core regions of the Mediter-
ranean through the disposal ofits suiplus grain. It simply lacked the
organisational apparatus necessary to make it respond quickly and
effectively to market developments acrossthe empire. It is telling
that in the case of the food crisis in Syrian Antioch, treated above,
there was no "marketing" agency which ensured that grain held
on neighbouring imperial estates and in Alexandriawas shipped
to the city in need. That only happenedat the emperor's personal
intervention. "9
Indirectly, of course, the state did promote some stabilisation
of market conditions across the Empire. The development of a
118 Josephus, Ant. Jud. 15. 307. Gamsey 1988, 255-257 emphasises how commercial
considerations would generally have been overruled by various political motives.
Market response in the Alexandrian export market would therefore habitually have
been heavily distorted.
' '9 SeeGrenier 1997for a relatedpoint. However,hegoesfurtherandseesstate involve-
ment in the grain supply ofRome as increasing uncertainty by making the market sector
muchsmallerandhencemoresusceptibleto shocks.Apartfrom suddeninterventions,
e.g. fixing a price level far below the market rate, I think this is going too far. The
presenceofenormous statesupplies will simply havereducedthestrategicbargaining
power ot private grain. Therefore it must have exercised a stabilising influence on
the situation. However, this is not to deny that there were still occasional attempts to
destabilise the market; sometimes they surface in the sources. But the contrast with
the republican period is marked. Not least the risk involved in speculation must have
increased in the imperial period. The executive powers of the emperor were much
largerand aboveall faster-workingthan thoseofthe republican state. He wouldhave
been able (and expected) to intervene more quickly if the situationthreatened to get
out of hand.
179
 IMPERIAL BAZAAR
tributary empire extending over a number ofdiverse areasrequired
some kind of administrative framework. Both the Mughal and
Roman Empires gradually created increased monetary and legal
cohesion between different regions. They also spread the use of
coins and written forms of contract, in the Roman Empire, par-
ticularly in the west, which in many regions had only used such
instrumentsrudimentarilybeforetheconquest. 120Thoughnottobe
belittled, I doubt that greater unity of law and coinage would have
beenanythingmore thanfacilitativein allowingmoreexchangeto
take place between localities. By themselves these were passive
media. Roman commercial law, for instance, was mostly reactive.
It did not so much drive institutional developments as develop
in response to the practices emerging among traders. Law and
coinage were part of the infrastructure. As such their creation did
provide important support for and expand the sphere of commer-
cial activities; butthey didnotconstitute organisations whichcould
actively coordinate andremould links between individual market-
places. 121 Moreover, these institutions were not primarily meant
to create unity between markets. They were designed to serve the
needs of the imperial state: to ease the collection of taxes, allow
it to make payments and regulate relations among its most impor-
tant subjects. On special occasions that might be taken to imply a
complete imposition of the institutions of the conquerors. "None
of the cities should be allowed to have its own separate comage
or system of weights and measures; they should all be required to
use ours", as the senator and historian Dio Cassius makes Mae-
cenas proclaim in an advisory speech addressed to Augustus. '22
That was a statement of ideology. In both the Mughal and Roman
Empires, though with variations in actual outcome and degree, the
needs of the state fell far short of the creation of absolute and
'2" SeeLoCascio20000,chapter i foraninterestingif perhapstoooptimistictreatmentot
this. Thesameholds for Meyer2004,chapter7, treatingtheuseofRomantabletsas a
form ofdocumentationin theprovinces. Thevery fewfindsoftabletsin theprovinces
do not warranther wide-rangingconclusions. See Aubert 2004on the use of writing
among merchants.
121 Morley 2007, chapter 4 is right to place the emphasis in his interpretation of Roman
law and coinage on their providing some basic state services (see p. 66 for the mostly
reactive nature of Roman commercial law).
122 Dio Cassius 52, 30, 9.
i8o
A ROUGH TRADING WORLD
immediate institutional homogeneity. 123 Pragmatism was crucial
to the operationofempire.
Thetendency towardshomogenisation wasgreatestinthesphere
of coinage. With the consolidation of Mughal rule under the
emperor Akbar (r. 1556-1605), an imperial currency was intro-
ducedwithgold, silver andcoppercoins. '24Localcoinages were
discontinued; no new issues were struck. In some regFons, for
instance the commercially important province of Gujarat, the old
currencyremainedincirculationfora considerabletimenexttothe
Mughal issues Butbyandlarge, theimperial currency managed to
establish aneffective monopoly inside therealm within a surpris-
ingly short span of time. The price the Great Mughal had to*pay
for this success waseffectively to relinquish the ability to manip-
ulate the coinage or pursue an active monetary policy. Instead the
Mughal state made the coins attractive by striking them at almost
100% purity and by keeping the monetary system open; anyone
with gold or silver could have it stmck for a fee at"the Mughal
mint. The value ofcoins, therefore, was effectively identical with
their precious metal contents.
The Roman state wasunwilling or unable to goquite sofar. The
greatconsolidator ofRoman rule, Augustus, alsoreformed thecur-
rency after the tumultuous civil wars of the late Republic. '^ He
created a three-tiered system basedon gold, silver and lower value
brass/bronze coins and fixed the exchange rate between the differ-
entdenominations bydecree.InthewesternpartoftheEmpire,the
Roman experience resembles the Mughal achievement. The virtual
purity of the golden aureus and the silver denarius secured their
acceptance and dominance from the beginning. By the reign of
Claudius,thecitiesofthewesternprovincesevenstoppedstriking
their own base denominations. In the eastern part of the Empire,
the situation was very different for silver and bronze. Here the
^ Lo Cascio20000: i8, "L'evoluzioneamministrativaaccompagnae talorafavorisce
questoprocesso diintegrazione, chee ancheunprocessodiprogressiva uniformiwone
e omogeneizwwne"(my emphasis). This is to go too far.
H.a^lLI 9"',i32,T435 for thT asic facts- For fl111 treatment of the Mughal monetary
a
12, '^s.tem'. seeHaide'', I9j?7a"dthevariouscontributions in Richards 1987.
"' My..treatment. oftheRomanimperialcoinagesystemis basedonCrawford F986b;
Howgego 1992, 1^94 and 1995, chapter 3-(wi'th interesting comparisons"to other
ancient empires); Harl 1996, chapters 4 and5; andWalters 1999,chapter 6.
i8l
 IMPERIAL BAZAAR
conquering power hadencountered the well-established monetary
practices of the Hellenistic states. To a large extent, the Roman
administration was satisfied with continuing this system, partly
leaving it to provincial communities, in groups or individually to
strike theircoinage. Thus locally minted copper coinageofsmaller
denominations continued and proliferated throughout the eastern
part of the Empire, reaching an unprecedented peak in the num-
ber ofissuing communities during thereign of Septimius Severus
(AD 193-211). A number ofregional silver coinages, though now
mostly issued by the imperial authorities, were maintained for a
longtime,inAsiaintothesecondcentury, inCappadociaandSyria
into the third and Egypt only saw the abolition of her own closed-
currency system under Diocletian (AD 284-305). 126 All this is
not to deny that even considerable monetary transfers happened
between the provinces. As is well known from later historical
epochs, such transfers are possible without homogeneity and if
the model presented in Chapter 2 is basically correct, those must
have taken place. The aureus would probably have been the main
medium for this; credit may also have played a role. '27 But the
idea of the imperial monetary system as removing all obstacles
to an unmediated transfer of economic resources between differ-
ent areas needs serious qualification. Unity of currency only came
about through a slow and gradual process. The fact that some of
the most important and prosperous provinces retained valuable
126 See Christiansen 1988 and 2004 for discussion of the Roman administration of the
Egyptianclosed-currencysystem. Nero seemsto haveexploitedthis to debasethe
Egyptiancurrency substantially, usingtheprofitstofillhisemptycoffersinRomeand
incidentallyto createa fiduciaryEgyptiancoinage.On a smallerscale,the Roman
state s. eemstohavebeenablealsotostriketheothereasternsilvercoinswitha slightly
lowerfinenessthanthemainstreamimperialdenarius(Harl 1996:98-106).Tolerance
in the east may, therefore, not just have been a sign of weakness. Compared to the
Mughal state, the Roman administration wasableto pursue a farmore active monetary
policy.
127 Hopk'ins 1995-6fortheimportanceoftheaureusin interregionaltransfers.Duncan-
Jones l994:"l67-170 and 2003 suggests, on the basis of hypothetical extrapolation
however,that the value of gold coins constituted a very significantpan of the total
imperialcoinage.SeefurtherHopkins1980contraDuncan-Jones1990,chapter2 for a
debateabout whether circulation ofimperial silver coinagehappened across theEmpire
orin regionalcells. Howgego1994offersaninterestingqualificationofthisdebate.
The two phenomenawere not mutually exclusive. Significantmonetary exchanges
couldtakeplacebetweenprovinceswithouterasingall regionalpeculiaritiesin the
compositionof the coinpopulation.SeeHarris2006andRathbone2003forcredit
transfers; De Ligt 2002for theprevalence of cashoverpapertransactions.
l82
A ROUGH TRADING WORLD
regional silver currencies for most of our period does say some-
thing about the limits of institutional integration in the empire. 128
Proceeding to the legal sphere, we find that ambiguous unity
gives way to considerable institutional fragmentation. Again, the
Mughal comparison offers a useful yardstick. The Mughals were
Muslims conquering territories with a population predominantly
subscribing to the variegated cults and practices we call Hinduism.
As Muslim rulers the Timurids saw themselves as the upholders
of Islamic law, the shar'iya. This was the law of the mperial
administration andcourts, supplemented withedictsissuedbythe
emperor to tackle specific problems. Occasionally this gave rise to
conflicts when imperial wishes clashed with established Islamic
doctrine. However, the main institutional gap was in relation to the
largenon-Muslimpopulation.Theadministrationhad.forobvious
reasons, to respect the locally varying customs of Hindu law. and
largely allowed the Hindu population to conduct its own aclmin-
istration of justice. Thus within the Mughal realm several legal
systems co-existed, and occasionally blended or clashed. I29
TheRomanemperorsdidnothaveto accommodatea canonical
imperial law, at least not until well into late antiquity. They could
therefore very gradually gofurther in legal integration. But essen-
tially they were faced with the same problem asthe Timurid rulers
of north India. Roman law was primarily the law of Roman citi-
zens.Originally thisgroupdenotedtheconquerors. Butgradually
The services of money changersandassayerswere, in otherwords, not madecom-
pletely redundant by Roman hegemony. This is, in any case, a problematic notion in
a currencysystembasedonthevalueofpreciousmetals,asthe'Mughalcasereminds
us.Evenwithouttheneedforexchangingcurrencies,themoneychangerstillhadan
importantrole in assayingthegenuineness,qualityandweightofcoins,for instance
makinga discountfor older, worncoins. Apuleius Met. X,~9 showsus the needfor
a money changer in a transaction involving gold coins and Suet. Nero 44. 2 distin-
guishesbetweenfreshandoldcoins. Ingeneral seeHarl 1996, 113-122, 259-260 and
Andreau1999,36-37forthecontinuedneedformoneychangersandassayersunder
the Empire.
129Mughalhistoriographyhasnotbeenheavilypreoccupiedwiththelegalsystem(s)of
fhe.H"Trf' t>.ut. seeHintze 'j?97' I75-l78 andHasan2004, chapter6, documenting
in detail the interaction in Gujarat of the imperial court with local legal customs
and adjudicatory bodies. Not surprisingly in the light of the separation of India and
Pakistanafter independencefrom British rule, historianshave beenmore interested
in the religious dimension ofthe meeting of Islam and Hinduism and the degree to
whichindividualMughalrulersattemptedtoaccommodateHinduelementswithinthe
imperial umbrella, e.g. Hodgson 1974,59-99.
i83

IMPERIAL BAZAAR
underthe emperors the right ofRoman citizenship wasextended
toparticularly privilegedcommunities groupsandindividualsin
therprovinces. In that"waytheright ofcitizenship spread,bul. it
happened patchily and in an uneven Pattern-7he, develoPmCTt
was also far more a phenomenon of the west than the east in the
empire l3°ThusRomancitizensremaineda minority untilC&ra-
calTa(AD211-17)issuedtheconstitutio antoninianaandextended
this'righttomostofthepopulation. '3' Evenafterthatdate,how^
ever,thelegalsituation wasfarfromhomogeneous. Atthetimeof
conquest'foreigners, peregnni, had to a large extent been allowed
to'maintain their old laws, just as they were left in charge_of the
local administration ofjustice. Sobeneaththeumbrella ofRoman
imperiallawa patchworkoflocalandregionalpracticesandcus-^
toms'had" continued to exist. '-32 As Mitteis long ago was thejirst
to'observe, thatcouldnothavechangedfundamentally afterCara-
cafla's'edict. Itisa simple matter oflogistics: thenumber ofpeople
thoroughly steepedinthecomplexrulesofimperiallawwerefar
too few. '33 Neither was there any drastic attempt to discard sud^
denlyprovinciallawenbloc.Itwasexpectedtoliveon,recogmsed
as custom in a Roman law context. Thus we find Greek orators
appealing totheparticular customs ofcommunities longafterthe
dramaticextension of the citizenship, and occasionally it is possi^
ble to observe how local legal practices begingradually to blend
into andtransform Roman law in the provinces. 134
. 3» SeeSherwin-White 19732t'orthe spreadofRoman citizenship. ^ ^ ^ __. :".. :.
.
3- Thoughsignificant'pocketsofnon-citizenscontinuedtoexistthroughoutlate^^
^ss"^1 ^s^^7rnuana~^esl9s6^how^^part of ^ empi
munTttes'wKh Latin rights, aphenomenon
restricted the to western
re
mhAt'also'have'adopted Roman law. But this will have varied from community to
community, see Galsterer 1997. . .. " . _, c. "l. ":,
G'ais'terer"''!'986 "is" a'clear'exposition of this. See Jacques and Scheid
. 3^
A"aplte^6. 'See" also Lin'ton 1993, though he appears slightly too optimistic
:radualuniformity of imperial law. _ _ ^ ..,. ___,., ":"
.
33 Mitteis 1891,chapter6. Modrzejewski1970providesMUluminatmga^lys^
re'la'nonshTpbet'we'enRomanlawand Egyptiancustomsbothbeforeanda^rCara^a.
^lft, ^Garnsey'2004:'notmg thatAeGreek orator Menanderrf ^^a^(t^
R'uss'e"lTa'ndWilson', Oxford; [98"!)eulogisedthecontinuedexistenceoflocalcustoms
evenaftertheextensionofRomanlawtothemajorityofcities.
-34Th^mix^ofRomanlawwithlocalcustomscanforinstancebegaugedfromfce
fifth-'century Romano-Syrian lawbook. Mitteis 1891 remams theclassictreatment.
See'alsoTaubenschlagl'955andModrzejewski1970and 1990.
184
1
A ROUGH TRADING WORLD
In fact, the situation was even more complicated than this. Even
before the constitutio antoniniana the divisions between the sphere
of local and imperial law were never firmly settled. The bound-
aries were permeable. Provincial populations could and did some-
times avail themselves of Roman legal forms and brought their
quarrels before the imperial provincial authorities. '35 A group of
wooden tablets from the mid second century testifies to the spread
of Roman forms of contract in Dacia. 13 These documents contain
examples of provincial subjects availing themselves of legal forms
which in strict law were only available to Roman citizens. Yet,
as recently remarked by Eva Jakab, the legal niceties were not
crucial to the contracting parties. They treated the Roman formula
pragmatically; it provided documentation of their transactions. 137
A fairly recent find in a cave of some thirty-five documents dating
from AD 93-132 in the newly established province of Roman Ara-
bia reveal the same kind of pragmatic adaptation. 138 The dossier
of papers concern the (family and property) affairs of a Jewish
woman, Babatha, belonging to the village gentry in the province.
We find her, among other things, engaging in a number of fairly
prolonged conflicts over property which seem to have reached
stalemate. In that situation, the parties try to involve the Roman
authoritiesto tip thebalancein theirfavour. 139Firsttheopponentis
presented with a lawsuit at the governor's court. This was presum-
ably intended as a threat because shortly afterwards the opponent
is made a not very favourable offer to settle the dispute out of
court. 140 What has especially caught the attention of scholars is
the close adherence, in some of the documents, to the formulas
and procedures of Roman law and litigation. All this has caused
surprise since Jewswere in general thought to have settled affairs
in their own courts. In consequence, there has been a tendency to
IJW
about 135 Augustus' Fourth Edict for Cyrene (in de Visscher 1940) rules that if the parties
agreed to it they could have an all-Roman court. However, the Verrine orations of
Cicero should teach us not to put too much emphasis on the aspect of consent to the
use of Roman courts. In practice, the Roman governor must have had broad scope to
exercise discretionary powers.
136 CILIII, 2, pp. 921-960. I37 Jakab 1997:166-170.Cf. Meyer2004:182-183.
13 Lewis 1989(P. Yadin)haspublishedandeditedthetwenty-sixdocuments written in
Greek.
139 Cotton 1993: 102-ioy for instance.
140 The pattern can be seen in P. Yadin 14 and 15 and in P. Yadin 23 and 24.
185
 T
IMPERIAL BAZAAR
make somewhat more optimistic assertions as to the unity of the
legal system in the Empire. 141
The evidence, however, does not so much support unity as
patchy integration, fluidity and regional variation. Jews in Galilee
are certainly known to have used both rabbinic and village/city
courts controlled by the Jewish gentry in addition to Roman courts.
In Egypt they, too, availed themselves both of Jewish and Hellenic
courts. I42 If the Babatha texts demonstrate a use of Roman law to a
surprising extent, they also contain testimony of major deviations,
pointing to a continuation of local practice, such as polygamy and
payment of taxes in pre-Roman coinage. '43 In one complaint to
the governor over her son's guardians Babatha suggests she will
be able to invest her son's trust fund at a rate of 1. 5% per month.
This probably implied lending the money out on interest. Thus
there is a good possibility that Babatha in her complaint to the
governor effectively suggested violating the normal Roman legal
annual maximum limit on interest of 12%. 44 Roman law was
appropriated by provincial subjects and adjusted to fit their own
purposes and cultural contexts.
To estimate the level of penetration of Roman legal procedures it
is alsonecessaryto takethe general conditionsoftheprovince con-
cemed into account. Both Dacia and Arabia lacked polls institu-
tions prior to the Roman conquest. This would make a more direct
transfer of Roman institutions more likely. But other provinces
complained against the use of Roman courts and defended their
141 See Johnston 1999: 9-1 I for a cautious, balanced attempt to formulate that position.
Other examples are Norr 1998 and Serrao 2000. Based especially on the Lex Jrnitana
(Gonzales 1986), Hackl 1997 argues this specifically for court procedures, with Egypt
alone as an exception. I find his conclusion unwarranted. It simply disregards the
contrary evidence and fails to take account of differences in status between provinces.
142 See Goodman 1983, chapter 10 for the co-existenceofrabbinic,Jewish-Hellenicand
Roman law courts in Galilee.
143 P. Yadin 16 is a census declaration to the Roman authorities testifying continuity in
money, measures and taxes. P. Yadin 26 documents potygamy. Wolff 1980:797-804
rightly distinguishesbetweenadherenceto Romanprocedureandcontinuationoflocal
custom in substance.
144 P. Yaclin 15. In his commentary Lewis suggests that Babatha must have been thinking
of lending the money out at interest. But he makes no connection to the maximum
of 12%. P. Yadin I I documents a loan advanced by a Roman centurion charging the
legally prescribed maximum of I % per month, so it was used by Romans in the area.
But we should not be surprised if the suggestion made here is true. Mitteis 1891:
156-158provides several similar examples.
l86
A ROUGH TRADING WORLD
legal "independence", asit were. '45Egypt, wherewritten records
and notaries were widespread before the annexation, sawthe con-
tinuation ofmuch ofherHellenistic legalpractices withrelatively
little Roman influence, particularly in the field of commercial
lawI4t) Newly enfranchised Romans in the provinces normalTy
alsoseemtohaveretainedtherighttousetheforeignlawsoftheir
hom^community. '47 This wasnot anempty gesture; a papyrus
from Egypt showstwo Roman citizens entering a contractuafrela-
tionshipwitheachotherusinga Greekformula.[48EveninItaly
wecanfindRoman citizens contracting withperegrini usingele-
ments of non-Roman legal forms. Among the wax tablets found
outside Pompeii in the archive of the Sulpicii bank is a document
pertaining to a contract for freight between the shipper Menelaos
from Ceramos in CariaandtheRoman merchant P.Attius Severus.
In thedocument Menelaos, writing in Greek, declares thathehas
receiveda sumofmoney(thesumrepresentingthecargo)fromthe
slaveofP.Attius Severustobereturnedaccordingtothecontract
forfreight. Onthenextpagefollows a surety declaration"by an
M BarbatusCeler,nowwrittenbya scribeinLatin.Formally,the
document complies with the standards of Roman law. BuUtafso
reveals interesting deviations from the pattern otherwise attested
in the Sulpicii archive. Both declarations conspicuously omit to
presenttheobligation asarisingoutofa mutually expressedagree-
ment aboutthecontracting matter. Instead, theyfollow the Greek
145T7hlcyrene Edirts, ofAugustus provideevicienceofthis(especially no.s.i and4,inde
WO).Theyarea responsetocomplaintsfromGreekprovincials'tlTa't'
to Roman Judges- Tlxrefore'the
are^°"s.i" o^, thin 8SLbemg sul'mltted emperor
'>ullt hat. ^omplaiTinvolvi",g from dlfferent cTfes mCy'renethey^ha'IL'a's
Greeks
^e;nhav. e, lrcek ^Idges;'lnd_SI!OUId neverbe Judged-
ag""-'"^r wiil"by a'P'urel'y
Roma'!.JUrypaneLFortheconti"ueduseofGreek'lawtoreg"ulate'relaU'ons'evet'n"at^
p^^aneve}among thed[ffeKMcit'eslseeMarsha"'98o:Anotherexampl^rfth^
continuinguseoflocallawisAthens,cf. Lintott 1993: 158.
'4b Modrzejewski1970;Taubenschlag1955.
147
ThLCIasslc.example is the Tabula ~Ba^ana. Here tribal chieftain is made Roman
a a
c^n/ahL'ure^e"lis. (cf' sherwm-white i973b). Interestmg"matenaTfo, :'c^r
as,prctsT of.;hi,s. problem is, Prov'decl .^ A"gustus'"Third Cyrene°Edict'(de vS.s^cher
to9^ IIruledthatenfra"chiseciProvi"cialswouldnotescape'the'dutie'soTt'heir'hon'ie
HRAm: V,9,\therontract wasfwPi8nusandtheformwastheun-RomansynchorSsis.
se^LFi^. "!LI2° wuh thejid'tor's comme"t:-"mufaum'inter'^verR ^^^
adihirogra^hIfoT amredJictam"-CAtr^r^A"m. too'>'warnot'Tstandard7o^''^
contractin Romanlaw,cf. GaiusInst.3. 133-134.
l87
 IMPERIAL BAZAAR
practice where contractual obligations are incurred by unilateral
declaration: unprodotto ellenistico-romano. 149 Roman law did not
create a homogeneousregime, it provideda flexibleframework. 150
This impression of mixture and fluidity is parallelled in courts
presided over by Roman magistrates in the provinces. These will
not always only have followed the prescriptions of Roman law.
From Spainthere is a very early illustration of this. The governor
conducts the proceedings in a dispute over boundaries between
two Spanish towns. He also defines the issue to the native jury
according to Roman notions. But the jury is asked to reach a
decision based on Spanish laws or customs!^' Several centuries
later and from the opposite end of the empire a summary of some
Egyptiancourt proceedings gives the same impression. A conflict
has arisen between a certain Dionysiaand her father. They take it
to the Roman court. During the trial, the father suddenly attempts
to close the question once and for all by claiming that according to
Egyptianlawthefatherwasentitled, atanytimehesawfit, to break
up the marriage of his daughters and take them back into custody.
The interesting thing here is the appeal to non-Roman laws in a
Roman court. In this particular case, the magistrate decided not to
condone the father's wish, probably on the basis of the precedent
of a denial by another magistrate in a similar case. But we have
several other instances where decisions are made expressly with
reference to Egyptian custom. '52
To sum up: there is no doubt that Roman law created a very grad-
ual, piecemeal and patchy increase in legal unity across the empire
during the Principate. But it never broke the pattern of strong
local variations in customs and practices. 153 Indeed, implement-
ing Roman law in the provinces required both pragmatism and
flexibility. Even when Roman law was employed in some form
149 TPSulp. 78 with the commentary by Camodeca. 15° Jakab 2006.
151 This is the so-calledTabulaContrebiensis. Cf. Galsterer 1986:22.
5 P. Oxy. 237: For an interesting discussion see Katzoff 1972. The interesting thing
about the precedents cited (P. Oxy. 2, 237, col. vii, 19-43) is that the decisionis left to
the wife. So in a way Roman law does not really enter the picture. P. Oxy. 2, 237, col
viii 21-43 cites precedents in other matters where the Roman authorities decided to
uphold the local law traditions. For the co-existence of many layers of legal tradition
in Egypt in general, see Modrzejewski 1970.
153 Galsterer 1986 and Gamsey 2004.
i88
A ROUGH TRADING WORLD
or other, we should expect considerable variations in interpreta-
tion from region to region. It was standard legal advice to Roman
administrators to allow local conditions to influence the outcome
of trials. '-54 That made sense. Instruments of legal enforcement
were not developed aselaborately asthe imperial law. Essentially
Roman law left it to private initiative to execute the rulings of
the court in civil law cases. '-5-5In that situation, as was alscHater
realised by Mughal judges, it would often be crucial to accom-
modate provincial sentiments andsecure the backingofthe local
community. 's6 Such considerations are even more pertinent in
evaluating cases where locals, such as the magistrates andjurors
in the small Spanish town with Latin rights, Imi, most of whom
would have had only a rudimentary knowledge of Roman law.
begantoconducttheirlegalbusinessaccordingtoitscomplicated
rules; it goes without saying that results must~at times have been
idiosyncratic, atleastfora fairly longperiod.'57Thusthesituation
was above all characterised by some institutional fluidity caused
bytheco-existence ofimperiallawwithmanylocally orregionally
varying practices. '58
An illustration of how the co-existence and clash of different
legal practices might attimes have caused uncertainty andconfu-
sion inthetrading world is provided by a letter written in early
medieval Cairo wherea Jewishmerchant community hadto han-
die both Jewishand Muslim law: "So-and-so confided to me one
anda thirdqintars ofindigo...This wasdonewithoutanyspec-
ification. I did not know whetherthe merchandise was given to
me as a commenda or a shipment, and if it was a commenda.
54 severalinstancescanbefoundintheDigest,e.g.25.4. 1. 15.
55 Nippel1995:iod-112.
156 SeeHasan^2004, chapters 5 and6 forsome casestudies ofMughal Gujarat showin(
holthe.a.dminis,tratio.", of. the shar"m Iaw ofthe emp"'ewas'°often~adjusted"or'
local predilections andpower relationships.
'" LexmhaMtantsofthIS
Irniuma,chapter93(Gonzales 1986)dictates"fcuseofRomancivil lawamons
the. sma11 Latin municiPiu>"- All the confidencem
strong""R^man6
;sat!on'Lth.is. newly fbund law has inspired needs to be tempered by recogmsmg'that
aalinstructionisnotthesameaspractice.JacquesandScheid1990:232notenst
at even the prescription to use Roman law still seems to expect~the~inhabUa'nt's"to
complement it with local practices.
SeeJakab 1997, 165-220forexamples oftheintermingling oflocal and Roman law
regardingthesaleofslaves.ThespreadofRomanlaw~wasaccompanied'byflexible
adaptation and adjustment.
189
 IMPERIAL BAZAAR
whether it wasaccording to Muslim or to Jewishlaw. "159Intimes
of disagreement such questions could be crucial. A ruling in a
casebytheemperorsSeverusandAntoninus(Caracalla)fromAD
204 reveals a clash of Roman and Jewish commercial customs.
Roman law espoused theprinciple that a stolen piece ofproperty
could alwaysbe reclaimed from thecurrent possessor by the on-
ginalowner.Thisheldirrespectively ofhowmanytimestheobject
had been sold or resold or of whether it had been bought in good
faith. Among legal scholars, incidentally, this state of affairs has
generallybeenconsideredanimpedimenttotrade.Itwasdifficult
for buyers to be surethatthere wasnotheft in thepasthistory of
goodsandhenceto beassuredofownership Not all aspectsof
Roman law worked automatically to the benefit oftrade, it is well
to remember. But this is not what interests us here. Jewish law
took a more even-handed view ofthe legal problem. It recognised
the right of the original possessor to vindicate his property, but
was also concerned to protect the innocent buyer. The good was
only to bereturned against payment ofthe sum for which it had
beenbought.Inthecaseathand,a groupofJewishmerchantshad
attempted toassertthisprinciple whensomeoftheirmerchandise
hadbeenrecognised bytheoriginal owner asstolen. Theemperors,
however, wereunyielding in imposing theRomanlaw view. "See
to it. therefore, henceforth to be more careful in your dealings, lest
you may incur not only losses ofthisnature butalsothe suspicion
ofcriminal activity", themerchants werereprimanded.'
The world ofbazaarsin the tributary empire
Imperial coinage andlawdid undoubtedly facilitate market trans-
actions, especially bybringingcoinageandwrittenlawtoareasthat
. 59 QuotedfromGoitein 1967,183.Inhisanalysisoflegalconditionsm theMoroccanSuq,
Chfford'Geertz 1979, 192-197, too, offers a comparative example oftheco-existence
andmixture oflocal, Muslim andFrenchcoloniallaw. _ . .. ..
.
60 C^VI, 2, 2:"Impp. SeverusetAntoninusAA. negotiatoribus Incivilemremdesiaerans,
ut'agnftas'res"furtivasnon prius reddatis, quam pretium fuerit
solutuma
doming
Cui:ate"i~gitur'cautius negoti'ari,
ne non tantum in damna
huiys modi, s^d^^lam^n
crimmis'suspicionemincidatis." Thecasewasingeniously analysedbyDaube^o^
7','whoTe'translation-has been used above. See further Yaron 1964 for ex amples
rfdiffenng'commercialprinciplesinRomanandJewishlaw.Ingeneral,seeGoodman
1983, chapter 10.
190
A ROUGH TRADING WORLD
had only a rudimentary pre-conquest experience of those instru-
ments. But it would be far too modernising to depict them as
providing a unifiedinstitutional foundationfor trade in the Roman
world. The Empire did not create a homogeneous and solidly gen-
eralised market sphere; institutionally, considerable heterogene-
ity remained. Some enduring residual fragmentation of law and
coinage in the empire, however, is only one aspect of the question
of the continuous vitality of local and regional trading practices.
But it doesprovide a good starting point, even ifwearenever likely
to see more than the tip of the iceberg. Gaius, in his treatment of
contracts in the Institutes (a manual of Roman law) includes some
observations on the differences between Roman and Greek forms
of contract. 161 This can probably be taken as a sign that the cul-
tural divide between east and west in the empire also found an
expressionintheorganisationoftrade.InthepapyrifoundatDura
Europos contracts composed accordingtoGreeklawcontinue tobe
predominant for a considerablenumberof years afterthe Roman
conquest. The same goes for Egypt, where Roman influence in
commercial law was very modest - in fact considerably less so
than in family law- and a local variety ofGreek forms ofcontract
prevailed. 162 Talmudic literature combined with various Roman
rulings (one of them is cited above) indicate the existence of par-
ticular Jewish-Semitic forms of commercial practice. F63 Various
scattered comments in the ancient sources also give an impression
of the existence of different customs for conduct at sea and the
ownership of wrecked goods washed up on the coasts. We have,
for instance, references by name to both Rhodian and Pontic sea
laws.I64
But tradingpractices includemore thanformal laws. Ina ruling
of the emperors Antoninus and Verus on buying and selling, the
" Gaius Inst. 3. 93-141 (especially 133-134). OnGreekcommercial lawseegenerally
Pringsheim 1950.
162 SeeTaubenschlag 1955:50onthestrengthoflocalcommercial lawinEgypt;Wolff
et at. 1978for a detaileddiscussionofthevariousformsofcontractin HeUenisticand
Roman Egypt; BradfordWelles, FinkandGilliam 1959:20-22 onthepapyri contracts
of DuraEuropos.
163 Yaron 1964.
164 Rouge1966,part3,chaptersI and5 offersa goodsurveyoftheevidence, evenifI do
not share his conclusions in all details.
i9i

IMPERIAL BAZAAR
judge is advisedto take into accountwhetherprice andmeasure
havebeenestablished in accordance withthe local custom. 105The
type of measure shows strong variations between regions. The
picture canbefilled out in greatestdetailfortheeasternpartof
the empire. The provinces of AsiaMinor, SyriaandEgypt each
hadtheir own predominant standards, none ofthem identical with
the main Roman measures. '66 Locally a regional standard might
varyquiteconsiderably.Inthepreservedpapyri,theEgyptiangrain
measure, theartaba,appearsinthreewidelyusedforms withminor
variations and a substantial number of additional local forms.'
A curious work by a fourth-century Cypriote bishop, Epiphanius,
addsto this imageby depicting a worldwitha bewildering variety
of measures and their variations. On Cyprus, for instance, two
kinds of medimnoi were in use. In the coastal area around Gaza
and Ashkalon a local wine measure, the sapation, could contain
from 14to 22 xestai.The xestes wasagaina very common measure
but known in many varieties. Epiphanius lists Italic, Alexandrian,
castrensis, Pontic'and Nicomedian types. '68 The picture can be
complicated further. Weighing andfilling is not anunambiguous,
objective and purely formal process. It involves many socially
negotiable choices.Therefore knowledge ofthe"formal" standard
was insufficient. The trader had to be familiar with the prevailing
measuring habits in a particular region. According to Epiphanius^
a metretes of wine on Cyprus contained 100 xestai, but would
normally be filled with 104 xestai from the press, the 4 being
discounted asdregs. Another example was whether a measure was
filled to the brim or with a heap on top. A gratuitous discount
mightalsooftenapplyasa matterofcourse,eitherinfavourofthe
buyer or seller according to theprevailing custom andbalance of
power. 169
.
66 Hari'1996':315-318offersa convenientoverview.Hultsch 1882,especiallypart5,
/idesa detailedsurvey ofdifferentstandards.
.
67 Duncan-Jones1979andRathbone19838providethebesttreatmentsofvariations m
-68 EDiohanTusDememurisetponderibus(ed.J.E.Dean1935),chapters590, 39, 41. See
K'ruit'andWorp1999forthebewilderingvarietyofEgyptianliquidmeasures.
.
(") EpiphamusDremensurisetponderibus, "chapters28and72b Seealsc'chapter2t^
adi^tmctKmbetweenfillingthemeasureeitherbypressingdownorshaking.P. Oxy.
192
A ROUGH TRADING WORLD
We must imagine the world of the Roman trader as steeped
in local and regional traditions. It would have been a regionally
fragmentedworld not quickly penetrated by the outsider. This was
partly a reflection of the great difficulties that stood in the way of
creating standardisation. Even when weightsclearly approach the
same standard, it is often possible to observe considerable devia-
tions; a cargo of lead ingots from the first century AD reveals fluc-
tuations in the Roman pound between 319 and 330 g.I7°Tampering
with weights seems to have been an endemic problem. Contracts
therefore often specify in detail the specific set of weights to be
used and where the weighingwas to take place, in order to ensure
some sort of control over the process. 17' The precise time for the
weighing or filling would often be of crucial importance. Most
products originated in agricultural processes requiring little or no
processing by modem standards. 172 Organic products are highly
susceptibleto change.Levelsofmoisture in a product,forinstance,
fluctuate with the season and distance from harvest. Such changes
greatly affectthe actual volume of a given product receivedby the
merchant. Products also show considerable geographical and inter-
annual variations, due to differences in growing conditions. Grain
is a good example. As Pliny informs us in his Natural History, a
measure of grain differed widely according to its region of origin
in terms of weight and baking capacity. Balearic wheat would,
for example, produce 35 pounds of bread to the modius as com-
pared with the Egyptian producing 20 pounds or less. 73 Qualities
2125 specifies that the grain has been measured with a total discount of 1. 5 artabai.
Duncan-Jones 1979 is excellent on all this. De Romanis 1998 is a recent, interesting
attempt, inspired by Kula 1986, to go beyond traditional metrology and combine the
analysis of variations in measure with social processes, even if too rationalising in
its attempt to fit the disparate information into an elaborate theory of changes in
bargaining power in the levy of customs on the Egyptian trade on the Indian Ocean.
170 Based on the table on p. 16 in Domergue and Liou 1997. Such small deviations add
up. In the Cabrera 6 wreck, a lead ingot of 32-33 kg might be up to almost i kg under
weight as a result of the minor deviations (no. 14 in the table). Duncan-Jones 1982,
appendix, conducts similar tests on weights found in Italy.
171 Falsificationof weights, e.g. Dig. 19. 1.32. Control of weighingprocess and specifica-
tion of measures, P. Yadin 21-22; P. Mich. 245.
172 Hopkins 19786: 50-52 is a rare attempt to take this aspect into consideration.
173 Plin. Nat. XVIII, 67. See also Plin. Nat. VIII, 189-197 for geographical variations
in wool and 19, 9-15 for flax. Wine of course is the classic example of a regionally
varying commodity.
193
 T
IMPERIAL BAZAAR
were difficult to control and prone to variations. It was not with-
out reason that Roman lawyers discussed how wine that had gone
sour should be treated in sales law. 174 As a consequence of this
general condition, adulteration ofgoods wasendemic. Wheat, for
instance might be mixed with small quantities of barley and earth,
sometimes even with grain which had gone rotten in the store-
houses. Expensive spices or frankincense could be blended with
lookalike substitutes etc. Pliny again offers a sobering catalogue
of methods. 75
The merchant had to keep his eyes open and be familiar with
every trick of the trade. In this respect close knowledge of local
conditions was crucial. For most areas of the empire we cannot
reconstruct regional practices in any detail or assess the degree
of fragmentation. But the distribution and production patterns
revealed by some of the major amphorae types may serve as a
pointer. Truly global productions arenotfound; amphorae typolo-
giesdiffer distinctly betweeneastandwest. Bothproduction and
distribution seem to a large extent dominated by regional and
in a few cases superregional patterns. '76 Dressel 6, for instance,
was firmly based in the Adriatic with exports finding an outlet in
Rome. and to a smaller extent in the nearby Aegean.Dressel 2-4
was linked with the coastal regions of the north-westem Mediter-
ranean (Tyrrhenian Italy, Provence andeastern Spain). Thesecon-
tainers, to a larger orsmaller extent, all entered long- ormore often
mid-rangenetworksofexchange.Buttheyneverlosttheirregional
character. 177 It is telling that the Baetican oil amphora, Dressel 20,
wasneverassimilatedto thedistincttypesofAfricaProconsularis,
eventhoughthelatterhada muchhighercarrying capacity (oil per
'74 Dig. 18.9. 2.
'75 SeeP\m. Nat. book XII. On adulteration of wheat, see e. g. P. Oxy. 708 and 2125.
eventheemperorrefrainedfrom iiuchpractices. CodexTheodosianus XI, 14, I orders
freshwheattobemixedwithstalegrainfromthestategranariesbeforedistribution.
.
76 ThecatalogueoftypesinPeacockandWilliams1986clearlyrevealsa wryregional
productionpatternforamphorae.Formoredetail,seeEmpereurandGarlan1986,
Amphores romaines 1989-,'Whitbread 1995, Eiring and Lund2004. Kruit andWorp
1999convincingly confirmtheregionalcharacteroftransportcontainersfromanalysis
of papyn.
177 Thisis"nottodenytheexistenceofcross-regionalfertilisationandborrowings &
Rhodianamphorae,influencingproductionstylesinItalyortheproductionofDressel
2-4 in Egypt. On the latter, see Empereur 1999.
194
A ROUGH TRADING WORLD
kg clay). This suggests that we should model the trading world,
not as a generalised market sphere, but as weak and patchily inte-
grated system consistingof a numberof segmentedandcompart-
mentalised circuits, some of only regional extent, others crossing
a number of regions. 178
The world of the Roman trader was an uneven, rough and het-
erogeneous place. Imbalances, asymmetries and bottlenecks in
transport, goods, informationandsocial institutionalisation, were a
chronicfeature.Thismadeforrelativelylowtransparencyandhigh
unpreclictability. It was a high-risk, high transaction-cost environ-
ment. I79 Correspondingly, it wascharacterised by large numbers of
relatively small traders andconsignmentsofgoods. Betternot put
all your eggs in one basket; a typical, medium-sizedship's cargo
of say 50-75 tonnes contained a considerable mixture of prod-
ucts in relatively small quantities. There would also normally be
morethanonemerchantbehindsucha shipment.Frominscriptions
found on the remains of cargoes in ancient shipwrecks, it is occa-
sionally possible to gauge some minimum numbers. The Pecio
Gondolfo wreck, for instance, contained small consignments of
fish sauce belonging to at least four different merchants. Another
exampleistheSud-Lavezzi2. HerewefindthenametagofAppius
lunius Zethus on both a group of lead ingots and the ship's anchor.
Shipownerand merchant are here one andthe same. But inscrip-
tions on some 235 copper bars found in the wreck reveal that the
ship catered for other merchants as well, not unlike the Phoenician
merchant captain described in Heliodorus' Ethiopian Tales; the
captain owned the main part of the cargo in his charge. But the
ship catered for a number of other merchants as well. The captain,
so to speak, brought his own competitors along to market. '80
178 SeeDuncan-Jones 1990,chapter3,usingthedistribution patterns ofstampedterracotta
lamps for a similar argument.
179 On the scale set out in North 1990the Roman Empire would have hadthe institutional
weaknessesofmarkets in the leastsuccessfuldevelopingcountriesandmostcomplex
historicalsocietiesbeforethedevelopmentofcapitalismin theearlymodernEuropean
world-system.
'80 HeliodorusAethiopicaV, l3-l5and 19.ForPecioGondolfo,seeLiouandRodriguez
Almeida 2000. For Sud-Lavezzi 2, see Domergue 1998. On the basis of some 1, 200
ancient wrecks, Parker 1992 concludes that most had more than one kind of article
on board; and this is based only on the visible remains. Most organic material has
long since perished and must be included in our analysis to produce a fuller image.
195
 T
IMPERIAL BAZAAR
Students of traditional trade in the early modem Levant and
Indian Ocean have referred to this sort of sociological profile ofthe
marketplace asthe pedlar-market. 18' That hasbeen strongly chal-
lenged, with some justification. '82 Even among relatively small
merchants it is possible to detect significant differences. There is,
after all. a considerable contrast between the small trader riding
with his donkey bringing a few goods from village to village and a
merchant taking several tonnes of grain, wine or fish sauce to the
market of a major city. Some large merchants surface in the main
markets of that world, too. In the Mughal Empire, Surat, a nodal
point in the Indian Ocean trade, boasted a small group of great
merchant princes. The Roman Empire canparallel the experience,
for example with inscriptional evidence from the city of Palmyra,
which played a central role in the long-distance trade in Oriental
luxuries: "This statue of Marcus Ulpius Yarhai, son of Hairan son
of Abgaros, was erected in his honour by the caravan led by his
son Abgaros from Spasinou Charax because he helped it in every
possible way. " 3
Pointing out the existence of these differences within the uni-
verse of the pedlars may appear a trifling matter, easily brushed
aside. 184It is tempting only to focus onthe many market imperfec-
tions of this world and reduce the merchants to powerless atoms
in a seeming chaos being thrown from sideto sideby the elements
Thesmall scaleofmanyconsigmentsofgoodscharacteristicofpre-industrialtradeis
alsodocumentedbycustomreceiptsandaccountsfoundin RomanEgypt.Sijpesteijn
1987,appendix i and0. Kerenike.publishedby Bagnall,HelmsandVerhoogt2000
nos. 1-117, show goods respectively leaving and arriving in peasant villages and
passing the customs station in the coastal city of Berenike as part of long-distance
trade. In both cases, however, goods are shipped in small quantities which could be
carried by a few donkeys or camels.
.
81 Most brilliantly, but also most controversially, Steensgaard 19730, building on van
Leur 1955 andLane 1966,followedby DasGupta 1979.
182 Chaudhuri 1979 and Bayly 1983, chapter 11 are the best-developed critiques. But
neithervanLeur,norSteensgaardclaimedthattherewerenobigtradersinAsiaorthat
therewerenoattempts atlocalmonopolies. Theyalsonoted themanyimperfections of
earlymodernEuropeantrade.Steensgaard'spoint,however,wasthatthemonopolies
enjoyed bythecharteredcompanies paradoxically didnotdistort, butbegantostabilise
and integrate markets.
'^ InventairedesinscriptionsdePalmyreX, 107.SeeDasGupta1979forSurat;Tchemia
2000;andCurtis 1991: 152-158for stratificationamongRomantradersin wineand
fish sauce, garum.
184 AsFinley i985a:33 and 136andWhittaker1993,chapter 15:5 tendto.
196
A ROUGH TRADING WORLD
and occurrences beyond their influence. The observation of Clif-
ford Geertz placed at the beginning of this chapter epitomises the
dilemma. Even if the merchants lacked the social power to begin
radically to transform the prevailing irregularities, they cannot be
described as passive victims. The enterprising spirit and shrewd-
nessofthe traditional trader is proverbial. A central point in thenew
institutional economics is that high transaction costs do not only
create market imperfections, they also foster different approaches
to the market. They generate distinct patterns of behaviour and
give rise to commercial strategiesdesignedto cope with the many
uncertainties. This is why attention to social differences between
the merchants is important. They show the existence of a hierarchy
andthus of a particular social system replete with institutionalised
forms of behaviour and specialisation of functions. The markets
of traditional trade should not be seen in terms only of one of its
players, the pedlar, they constituted an entire social universe - the
bazaar.
I use the concept ofthe "bazaar"advisedly. As a type, the bazaar
describes the form oftrade characteristic ofcomplex agrarian soci-
eties; it isnotidenticalwiththeclusterofpetty retailersandface-to-
face bargaining experienced by tourists travelling the Middle East
today. These are only the sad, if sometimes still exotic and colour-
ful, remains of a formerly vibrant trading world - a backwater now
bypassed by the main currents of trade which have all been sub-
sumed underthe institutions ofmodem capitalism. Inpre-capitalist
societies,thebazaarrepresentsthemaincommercial system, in the
colonies of European empires, the indigenous sector. Historically,
long-distance and wholesale trade, brokers and auctions have all
formed part of commercial life in the bazaar. '85 Understood as a
system and a hierarchy, the bazaarhas involved much more than
a mere cluster of shops in an urban centre. It also reached out to
the periodic fairs and markets ofpeasant society, whose schedules,
as both the ancienthistorianandanthropologist can confirm, were
ls5 Failureto includesuchphenomenaintheanalysisisanimportantweaknessofGeertz
1963, largely remedied in 1979. For these activities in India, see Chaudhuri 1978 and
Bayty 1983, chapters 10 and 11. For Rome, see Talamanca 1955; Andreau 1974 and
1987; Rauh 1989; and Ankum 1972, who on p. 377 made the important observation
thatauctionsalesacquiredtheirimportancein Romansocietypreciselybecauseofthe
absence of many modern market institutions.
197
 7
IMPERIAL BAZAAR
frequently coordinated with market days in nearby cities. Stretch-
ins back to its rural hinterland, the bazaar also projected outwards
to enter widerurbannetworks of exchange.186But in comparison
with the capitalist institutions gradually evolving in seventeenth-
and eighteenth-century Europe, the bazaar represented a distinct
form oftrade. Institutionally it hadnot somuch begunto transform
the basic uncertainties of pre-industrial trade, as it had learned to
live with its asymmetries and irregularities. As a trading envi-
ronment, the bazaar is distinguished by low transparency. The
combinationof solidroots in peasanteconomies, sometrade over
longer distances and a considerable degree of uncertainty, makes
the bazaar an attractive model for the functioning of markets in the
Roman Empire.
Clifford Geertz has developed the bazaar as an ideal-type and
givenittheoreticalcoherence. ' 87Heidentifiesa numberofdefining
characteristics. Key among these arepoor information, fragmented
organisation and low standardisation: the features described above
as central for our understanding of Roman trade. In comparison
to modern markets, bazaars had few institutions collecting and
disseminatinggeneralisedinformation,suchas,for instance,price
statistics. TheRomanandMughaltradingworlddidnot knowthe
development, following inthe wake ofthe introduction oftheprint-
ingpress, ofwidely published price information innewspapers and
other media. '88 Instead merchants had to rely to a greater extent on
personal connections andcorrespondents. The bazaarwasa place
rife withrumours. In terms ofcommercial organisation,the bazaar
represents a fragmented environment. Prices differed, as wehave
seen, considerably, both inside and between markets and, as shall
be examined more fully in Chapter 5, business was situated within
a household context. Finally, there was significant variation in the
186 Geertz 1979: 188-189(for localaswellashigher-levelconnections);deLigt 1993is
nowfundamentalonfairsandruralmarketsin theRomanEmpire;Shaw 1984remains
valuable; further discussion ofthe Roman andCampanian evidence is provided by the
contributions of Andreau, Marino and Ziccardi in Lo Cascio 2000a.
187 The following ideal-type summarises the analysis anddescription of'Geertz 1979, two
key sections are pp. 124-125 and 197-217.
.
88 The practice of including among other omens the prices of central commodities in
astrologicalrecordsassignsofdivinefavourordisaffectioncouldhardlyhavefilled a
similar role and there is no evidence to suggest that, pace Andreau 1997.
198
A ROUGH TRADING WORLD
standards of goods and commercial customs. Taken together these
features put a high premium on the phenomenon which Geertz in
another context has deemed "local knowledge", close and detailed
understandingofconditionsin a specificlocale. As a consequence,
Geertz explains, success in commerce was less a question of gen-
erally assessing the situation across a wide number of markets
and then weighing your options than it was about getting it right
in your own particular market. "Sampling", of course, did take
place; but the coordination of activities in separate markets was
less important than in a more closely integrated system: What does
it help you that you can get a cheaper consignment of cloth in a
distant competing market, if you cannot be sure of its quality?
The widespreadhigglingandhaggling,whichmore than anything
else has come to symbolise the bazaar in common perceptions,
is a reflection of this problem. A prolonged period of intensive
negotiation enables the parties to examine and reach agreement
on all the intricate details and conditions relating to the particular
set of goods on offer. It is mastery of the specifics whichensures
success. 9
This has significant implications for the conduct of trade. The
bazaar should not be understood as a failed attempt to transform
the many uncertainties of its environment. Instead the pervasive
risks andimbalanceshadbeendomesticated,asit were. They were
to some extent reproduced and perpetuated by the institutions of
the bazaar. Rather than removing the endemic irregularities, the
merchants of the bazaarhaddeveloped commercial strategies that
allowed them to co-exist with the risks and often to turn them to
their advantage. 190A good illustrative example can be found in the
report, mentioned above, made by Francisco Pelsaert, the Dutch
factor stationed in Mughal Agra. Pelsaert recommended spread-
ing and coordinating activities between different markets in order
to be less dependent on the varying individual local situations
andcreategreatercontinuity. In contrast, the strategypursuedby a
189 See Geertz 1979: 221-229 for the logic of intensive bargaining as distinct from exten-
sive sampling.
190 VanLeur1955;Goitein 1967;Stoianovich 1974;Bayly 1983,chapters 10-11;Chaud-
huri 1985; Ray 1988; Yang 1998; and Datta 2000 provide a good impression of the
functioning of these kinds of market.
199
 IMPERIAL BAZAAR
small group ofPelsaert's Indian competitors sought to profit rather
from the irregularities ofthe indigotrade. Insteadofaimingat per-
manent control of the market situation, a number of Muslim and
Hindumerchants, residentin theproductionarea,boughta portion
of the indigo harvest from the peasants some months in advance,
sufficient to put them in a favourable position in years with a bad
harvest. As a result, they controlled the local market in years of
low yield and were able to pocket a temporary monopoly profit
in selling to the Dutch. In good years, however, the Dutch were
able to do without them. 19' Indian history is full of such, often
locally entrenched, groups of traders aiming to take advantage
of and increase the market imbalances in order to secure tempo-
rary, short-lived local monopolies. 192 They are not foreign to the
Roman world either. One thinks of the food crisis in Syrian Anti-
och. treated above; or that stock character of classical moralising,
the grain-speculating merchant. Often mistakenly portrayed asthe
quintessential capitalist relying on superior organisation to tie mar-
kets together, he represents the epitome ofbazaartrade, always on
the lookout to make a profit from irregular and excessively high
prices in some locality. '93 Instead of integrating markets closely,
the approach of the bazaar would often have been more oppor-
tunistic or speculative, with the merchant constantly seeking to
bring himself in a situation where he could benefit from a market
imbalance.
Such commercial opportunism was one of the strategies widely
employed by merchants to navigate therough waters ofthebazaar.
A second, no less important, aspect was the organisation of pro-
tection. If bazaartraders were not generally the beneficiaries of
"" Pelsaert 1925: I5-l6.
192 Datta 2000, especially 208-220, provides abundant evidence of speculative, destabil-
ising local monopoly strategies pursued by merchants in late Mughal/early colonial
Bengal. Datta, p. 325, correctly insists that these practices were crucial for the com-
mercialisation process in the first place.
"» Philostratus Vita Apollonii, 4, 32 offers an example of the grain speculator. SB 7242
(= Johnston 240), is a business letter seeking to use "insider-knowledge to create
a temporary local monopoly in peaches. Rouge 1966, part 3, chapter 6 mistakes
such activities for the capitalist speculator. See Kudlien 1994 for several valuable
observations on the particular mixture of freedom and monopolies which made up
ancient trade; Giarclina 1986 for an interesting analysis of the secretive mentality of
the ancient merchant.
200
A ROUGH TRADING WORLD
mercantilist-style privileges, they still hadto obtain a level ofpro-
tection from political authorities. Customs dues can, in fact, be
understood as a kind of protection money, with all the ambiguities
this implies. Chapter 4 will explore how the price of protection
was negotiated between merchants and the "protection selling"
authorities. How this relationship was worked out, provided one
ofthe underpinningsofcommerciallife. A thirdmechanismwhich
is of great importance to the bazaar merchant is the formation of
personal relations of trust and communal ties. These will be the
object of Chapter 5. If uncertainty is high, then one way to tackle
this is to forge specific alliances. You do not as a rule do business
with all merchants, you preferto trade withyour personal connec-
tions or members from yourowngroup, people whocanbetrusted
to a greaterdegreeandwhomay be willing to bailyou out in times
of trouble. Clientelisation and formation of group solidarities, as
Geertz explains, serve to cushion the insider againstuncertainty.
But this is achieved by making the market situation less trans-
parent and more impenetrable for the outsider. To the individual
agent, the market never presented itself as a generalised sphere
of commercial activity, a level playing field, it was only acessible
through a "limited set" of concrete relations. 194 Combined then,
the next two chapters will explore some of the distinct strategies
andmechanismswhichenabledRomanmerchantsto copewiththe
ubiquitous uncertaintiesandmarket asymmetries ofa bazaarenvi-
ronment. If I am alloweda metaphorfrom epidemiology,they will
try to show how uncertainty went from being virulent to endemic.
194 Geertz 1979:2l7-22o(clientelisationandmarketfragmentation); 188-189(themarket
as a specific set of concrete limited relations); 154-172 (communal organisations);
137-138 (the relationship between protection and passage toll as underpinning trade).
201

A THIN LINE: PORTORIUM, PROTECTION
AND PREDATION
Shouldyouwishto abusea taxfarmer, youmighttry saying: oppressor, burcten,
garrotter,raidingplunderer,shark,fiercerthanthesea,wilderthana winter^storm.
op pressor "of'thFdown-trodden, inhuman, insatiable^immodera^'
burdensome,
Je^ymoneygrubbingsod,violator,strangler,crusher burglar, st-ip-Jack-Mked,
snatche^thietD -overcharger, reckless, shameless, unblushing, painanimal,tten^c^.
^ageTwilcUnhospitable, brute, reef, sharp rock, wreckage, pure
theoAerviletermsyoucanfindtoapplytosomeone'scharacten
Pollux, Oiwmasticon, 1A, 30-3'
He[thetaxofficial]mustassessaccuratelythevalueofthegoodstha"heme^
will reveal any
chantro f"'those'mahals bring in so that no future investigation
dis"crepancy.-Therebytheofficersofthatplaceaccordingtothejaluatiol\m^
c'ollectt"th'eproperrevenuesfromthat(merchandise)fortheexaltedstate.Apart
fromVhees^biished rate, hemust not take anything from themerchants. ^ ^
"Richards. Ooc-umcnf Formsfor the Official Orders ofAppointment inthe
MitRhal Empire,
In his description of the route taken by the caravans^importing
frankincense'from southern Arabia to Gaza on the Mediterranean
littoral, Pliny observes how the costs of various local levies^
toUs accumulated to a grand total of 688 denariiper camel load
evenbeforecustoms wereexactedbyRomanpublicans.'
speaking, this amount wasequal to the annual salaryot^three
RomanTegionariesortheaveragepercapitaincomeofsometwelve
personsintheEmpire. 2 Thisisa powerful,ifadmittedlye^n^
reminder that one of the most important expenses in pre-modern
tradewasthepayment ofcustoms duties. Oftenitwouldhavebeen
' PlinyNat.XII,32.OntheArabiantrade,seeYoung200J,chapter3.Thetranslationof
PoTluxisheavilymodifiedfromMeijerandNijf1992:82.^ ^ ^^^ ^ ^ ^
^ ^leuoArd^aTy'^nua'rp ay"ofa"legionary around the time ofPlmy ^econd half^n;'^
SuOTydI^DTwTU2a2'5pdaeynaomaTch^Zn^^^^
a percapitaincomeofapproximately 57. 5 denarii(HS230).
202
A THIN LINE: PORTORIUM, PROTECTION AND PREDATION
far greater than or at least similar to the costs of transport. 3 In the
last chapter, for instance, we saw how exciseduties quite possibly
burdened even the low-value, high-bulk grain trade in the province
ot'Arcadiain late RomanEgypt as much as or more than transport
costs.
Attention is too easily focused on the high-profile phenomena of
pirates and bandits when assessing the problems of ancient mer-
chants. In everyday experience, the merchant would on average
have had to deal with customs officials much more frequently.
That could be bad enough. The Empire was not a Rechtsstaat. A
in
routine encounter with a customs official could be a harrowing
and all
experience. At times there would not have been much to set him
apart from the common bandit. As the rich vocabulary of Pollux,
the second-centuryauthorofa Greekthesaurus, showsthe distinc-
tion was in popular perception commonly seen as a question of
semantic power rather than of real facts.4 "There is no one who
does not know how great is the audacity and recklessness of the
bands of publicans", the king of Roman jurisprudence, Ulpian,
added.5 Imperial governments might on occasion, as illustrated by
the Mughal set of norms for good conduct issued to its customs
collectors, attempt to curb the worst excesses. But in practice it
was only a thin line separating protection and predation in the
trading world of the Roman and Mughal Empires. This chapter
will attempt to examine the process of customs collection as a
balance of power between state fiscalism, local elites and com-
mercial agents. That balancewas never primarily struck in favour
3 This point has been made most strongly by F. C. Lane 1966, passim, and later developed
by Steensgaard 19733, 60-68 and 1981. Rouge 1966, 450-451 followed by De Salvo
[992, 3I7-32I recognisethe importantpart playedby customsbutnevertheless speculate
that the technical costs of transport and handling would have been much higher. That
view is notjustified, as a general rule. A papyms may provide further illustration (BGU
Ill, 697). A camel driver hastransported alumfor merchants in the village of'Soknopaiou
Nesos. The distance covered is relatively short, yet he had to pass one customs station.
That already brought the payment of customs to halfthe transport costs.
4 It is interesting to note that accortfing to Ulpian it was generally agreed among Roman
lawyers that a wronged party could sue the customs collector both under the special title
regulating collection of tolls and under the title for simple theft and robbery, cf. Dig.
(Ulpian) 39. 4. 14.
5 Dig. 394. i2. pr. : "Quantae audaciae, quantae temeritatis sint publicanorum factiones,
nemo est qui nesciat."
203

IMPERIAL BAZAAR
ofthemerchant. The imperial peace, it will appear, may well have
benefited trade; but traders were also made to pay.
Predation
Abuseofpower,confiscations,racketeeringandoverchargingwere
frequently complained ofinconnection withtheexaction oftolls
Aninstructiveexampleisprovidedbyanedictissuedbytheprefect
ofEgypt in the micfsecond century AD: "I [the prefect ofEgypt]
aminformedthatthecustoms collectors haveemployed fraudulent
andclever tricks againstthose who are passing through the coun-
try andthatthey are,in addition, demanding whatisnotowingto
themandaredetainingthosewhoareinurgenthaste,inorderthat
some may pay for a speedier release. "6This is not a phenomenon
which was'unique to'the Roman world. It seems to havebeen a
general problem in pre-modern societies. Fromthemanytrading
records, descriptions andgovernmental orders surviving from the
MughalEmpirea similarpictureemerges. 7 Ina segmentary impe-
rial"formation, as discussed in Chapter 2, offices invested their
holders with a personal share, as it were, in state power. Offi-
cials could and did normally expect to profit from their entrusted
positions. Hence the widespread practice of state auctioning of
offices, rights andperquisites. Theprofit nexus wasneverclearer
than in the field of tax-farming where the state delegated part of
its sovereignty, the right to collect taxes, to private contractors
in return for a fixed sum of money or a proportion of the tax
proceeds. Most customs andexciseduties in theRoman Empire
were collected by tax-farmers or, to use another term, publicans
(publicani)s Naturally they would try to make their investment
6 P. Princ. 11, 20, col. l with Reinmuth 1936.
7 Hasan"2004,~34^40(revenuefarming)andchapter7 (customscollection)^ CteGupta
the Mughal . far
3', 'e. K"'Tl'6'4T and 206-207. For examples from sources,
see
man^
^iym^nare dicTs)"pres:erved'in e. g: chapters 105 and 117 of &eM^'at-[Ahmad^
th'e'descriptionby Tavernier 1889, vol. i, chapter 2 of the payment of customs in
8 DeLaet1949,followedby Vittinghof . 953:3»4-394andthelately inpartly France^OOK
43'3T3 8^^g"ed thaUarge parts"of the Empire during reigns o^Marc^sAyreJms
md Septimius Severus sawa transition to direct statecollection ofcustoms. Brunt 1990,
^ap"tert'l7"convincingly"demonstrated thattheepigraphic evidencefarfromsupports this
204
A THIN LINE: PORTORIUM, PROTECTION AND PREDATION
pay. Thus in the system of customs collection there was a built-in
propensity towards squeezing the merchants as much as possible.
Abuse was never just a question of the occasional corrupt official
falling prey to the temptations offered by a position of power; it
was structural. 9 Even in cities where commercial interests could
be expected to enjoy better representation than on average, con-
frontation between the collectors of customs and the traders was
inevitable. The surviving customs law ofPalmyra, the Roman cara-
van city in the Syrian desert, was precisely the product of such a
conflict. 10 Not that the relationship was always a matter of brute
force, confiscation or overt overcharging. The squeezing mecha-
nism could take many forms. The customs officials had several
more subtle tools at their disposal.
Often customs duties were levied at a percentage rate charged
on the value of the goods carried by the traders. This required
the customs collector to make a (subjective) judgement on their
value. That left plenty of discretionary scope to the publican for
increasing his intake by going for an estimate in the higher ranges
of the normal price band. Customary practices also played a con-
siderable role in collecting the duties. By appealing to tradition
publicans were often able stealthily to introduce a host of extra
administrative charges on the traffic passing through their sta-
tion. " Egyptian papyri show how, in addition to the actual tax, a
sometimes very significant number of surcharges applied as well.
interpretation. There is no evidence of a general shift to direct state collection. In any
case, if such a transition did occur, it never comprised the entire empire and it would have
been short-lived. From the beginningof the fourth century the evidence unequivocally
shows the predominance of tax-t'arming. A more likely interpretation would link the
changes observable in the late second-century epigraphic record to a gradual transition
in some provinces from the use of large tax-t'anner companies to a decentralised system
of more localised contractors. While response to Brunt has varied in the detail, his
broader picture seems very widely accepted, cf. Cottier 2003; France 2003; and Lo
Cascio 200311.
9 Pace the apologetic and unanalytical comment by Sijpestein 1987: 90-91: "That they
[the customs officials] mi.sbehaved on occasion is only human. " This simply fails to
take account of the social context of the phenomenon. The widespread use of auctions
to delegate the right of collection heavily favoured the interests of state and publicans
over those of the tax payer, as argued by Malmendier 2002 in her study of the auctions
for the Roman companies of publicani during the Republic.
1(1 Matthews 1984.
'' Overestimationof goods: (Ps)-Quint. Declam. 340 with Vittinghof1953:395-397;For
the role of custom, e. g. Dig. (Paul) 39.4.4. 2: "In omnibus vectigalibus fere consuetudo
spectari solet" and Codex Justinianus IV. 62. 4. Gradual introduction of extra charges,
205
 IMPERIAL BAZAAR
Here is an account dating from around the turn of the first century
AD for the payment ofthe Memphis harbour toll:'2
550artabaeofwheat,customs,44dr. ; wharfage(?), 6 dr. ; tax 4 dr. ; examination
dues. 4 dr.; to the banker, I dr. ; administration, 5 dr.; exchange, I d''-; to an
interpreter, ! dr. ; rudder (possibly a pilot's fee), 14 dr. 4 ob. ; a guard and boat,
4 ob. ; for the clerks. 4 dr. ; to Artemis, I dr. ; affidavit, I dr. ; receipt, 4 ob. ; on
accountof libationfee to examiner,2 dr. ; total, 91 dr.
It isnoteworthy howanextra47 dr. is addedto theproperduty of
44 dr., to cover various administrative charges, thus doubling the
amount to be paid. The surcharges would not always have been at
quite this level. Other accounts show them comprising a smaller,
though still significant, proportion. A shipment of some 400 wine
jarshad118dr.addedtotheexcisedutyproperof300dr includ-
ing 30 dr. in additional charges ("TTpoo-5iaypa(p6|j£va"). 13 In a
list covering a month's collection of duty andexamination fees at
a small customs stationin a village oftheFayumthefinal account
is drawn up by adding to the accumulated sum of 959 dr. 3. 5 ob.
a further 8. 5% for administration ("xsipi cmKov") and exchange
("aAAay-n")/4 Takentogether thepapyri provide telling evidence
of the considerable ability ofpublicans and administrative person-
neltomanipulate orinfluence theeffectiveratefortheexactionof
customs. '5
The imperial states did try to constrain the activities of the
customs collectors within certain bounds. Mughal and Roman
emperors alike occasionally reprimanded customs officials and
admonished them not to charge more than their due. But one
maylegitimately question howefficient theseinitiatives were. One
cannot quite escape the impression that most were rather irregu-
lar. remote, ineffective and recommendatory in character - to
Tac. Ann. XIII, 51. The Palmyrene tax law (C/S II, 3. 3913; translation in Matthews
1984),sectioni a (preamble)indicateshowtheflexibleframeworkofcustomgavense
to conflicts betweenmerchants and publicans.
.
2 P. 'Oxy. 1650, 001. i. Seeespeciall/deLaet1949:317-321forananalysisofthis a
related documents. ,....,."
13 P. Oxy. 1651. The editors wrongly published the papyrus as "account ot rreigm.
14 SB7365, dating to AD 104. A translation is provided byJohnston 1936,no 347.
'5 See'also'thee'xpressexemption,recordedbyBean 1954: 102,ofthemerchantsfrom
havingtopayasurtax"inthenameofAphroditeorasanykindof'declarationfee'
during a festival suspension ofthe regular customs duesin Caunos.
206
A THIN LINE: PORTORIUM, PROTECTION AND PREDATION
paraphraseFarhatHasan'srecentanalysisoftheMughalsystem. '6
The British and Dutch chartered companies trading in seventeenth-
century Mughal India had by imperial order been exempted from
all levies inland and would only have to pay the imperial duty
on exports and imports exacted at entry or exit to the Empire in
the coastal regions of Bengal and Gujarat. That was imperial law.
Reality proved quite different. The companies had repeatedly to
defend their exemption. Occasionally the Great Mughal would
issue a. farmcm (edict) enjoining customs officials to refrain from
exacting the illegal duties from the British and Dutch. That might
help, briefly. After a short while the old ways returned and the
companies again found themselves pressed to pay inland duties,
irrespective of the ruling of the Mughal court. I7
The story does not read very differently as told by Tacitus com-
menting on Nero's attempt to curb the most excessive abuses of
Roman publicans. At the end of a list of Nero's countermeasures
Tacitus makes the following sobering remark: "and there were
other very fair rulings which were observed for a time and then
allowed to lapse. The abolition, though, of the 'fortieth', 'fiftieth'
and the other illicit (sur)charges the publicans had invented, is
still in force. '"8 Considering the evidence we have just seen for
the collection of excise duties in Egypt, some of which was con-
temporary with Tacitus, it is clear that the modest success of the
Neronian reform did not consist in the eradication of the demand
for surcharges across the Roman Empire. If Tacitus is right, as he
may well be, the lasting influence of the reform would have to be
seen as limited to specific sets of customs, such as for instance
the Asian portorium. More probably, Tacitus' observation meant
no more than that the emperors periodically simply repeated the
ban in judging legal disputes only to see it lapse again during the
everyday practice of regular commercial life.
Cf. Hasan 2004: I 19: "Direct imperial intervention in the management of the fiscal
system did take place, but it was irregular, remote, recommendatory and usually inef-
fective."
17 See Hasan 2004: ll6-l [8 and 124-125 for a translation of afarman by the emperor
Aurangzebdating to 1667 regarding the trade of the companies from Surat in Gujarat.
Seealsothefannanfrom 1665containedin chapter 106ofthe Mir'at-iAhmadi(p. 235).
See Prakash 1972 on the precarious hold of the European companies on the privileges
they enjoyed in the trade from Bengal.
18 Tac. Ann. XL[I, 5l.
207

IMPERIAL BAZAAR
The impression of relatively feeble state constraints on cus-
toms collectors receives confirmation from the legal prescriptions
and rulings preserved mainly in the Digest. ^ The Roman jurists
introduced a special title/action to regulate the activities of pub-
licans. 20 The basic principle seems to have been that they were
liable for double the sum they hadunlawfully taken. As observed
by thejurists, the paradoxical result was that the tax collectors
were generally in a better position than normal persons illegally
appropriating'the property of others; they faced a penalty of the
quadruple amount Therefore thejurists do not seem entirely to
have excluded the possibility that a wronged party could choose
to sue for simple theft instead. 21 But that does seem more of a
theoretical option than a reflection of the day-to-day running of
things.Whyelseshouldthejurists everhavebotheredaboutdevel-
oping andcommenting upon thespecial title? A ruling, attributed
to the Severan jurist, "Paulus, tried to lessen the gap slightly by
beefingupthepenalty, ifviolencehadbeeninvolved, totriplethe
exacted amount with the added possibility of extraordinary pun-
ishments - the latter measure for the sake of"public discipline". 22
However, generally the current seems to have run in the opposite
direction. There was a broad consensus among the jurists that an
accused publican could free himself of all further liabilities^sim-
ply by returning the wrongfully exacted objects or money. Some
approach towards
'9 In the late Empire, from the time of Constantine,
the state abusive
'lTcans"doesl "formaUyseem to have become significantly inUnewith^
tougher,
raTtendencyin penalties.Failureto recogniseimmunitiesor overcharging^were
^o'wsometimesorderedtobepunishedbydeathorexile(C.Th.IV, 13, i and3;CJIV^
6l'^(^4). Theoldmorelenientprocedures,howeverwerestillincludedinthelegal
axlifications(a,dex~JustinianusIV,62,3;Dig.39.4?^"").Thusit isquestionable
whether the harsher imperial punishments ever translated into a coherent new regime
(d'e'Laet"1949,475-482alsoemphasisestheco-existenceofnewandoldmeasures^
Mor7'unpo'nanriy;'ifanythingtheneedtoresorttoextremeactions^uch_as/he^e,at
the limited
v"i7Drobablv rather an "indication than
of weakness strength -
jibllity
^"the'TO'v^rnment'to'regulate. daily practice. Capital punishment is not a very flexible
t'ooT.'Onbalance;it wouldprobablyhavebecomemoredifficultforpeopleofmoderate
sta'tus'to"obtaina judgementagainstquasi state officials suchas publicansbecause
ther>enaltv-was so"drastic and potentially much more disruptive to socialpeace.
persrons"ofinfluenceitwasa differentmatter. Publicanswouldnowhavefounditeven
more difficult to challenge such people.
20 DeLaet 1949:444-445 for a basic, though notvery analytical, treatment.
21 Dig. (Ulpian) 39. 4. 1 . 2-4. _ .. . ., . , __. _. collection
""". ".:, of
22 ('exce rpted from
D?g. 39'^9. '5 the a^late
SententiaePa^ third-century
rulings'ascribedto Paulus,butcertainlynotall ofwhichweregenuine).
208
A THIN LINE: PORTORIUM, PROTECTION AND PREDATION
maintained that this still applied even after a Roman magis-
trate had granted a trial against the culprit. 23 This, of course,
would have opened an avenue for a last-minute "amicable" set-
tlement where the obviously wronged merchant could have his
goods returned without having to go through with the trial. But
this has to be set against the much greater leeway such a mea-
sure granted the publicans in their negotiations with dissatisfied
merchants. The possibility of returning with impunity enabled
them to hold on to goods for longer, draw things out and test
the stamina, determination and social clout of the opposing mer-
chants. Given the seasonal nature of much trading, and the time
and effort required to seek out a court, sometimes far away, this
concession must have worked to the distinct advantage of the
publicans. 4
The relative leniency shown in the legal rules against the col-
lectors of customs duties is in stark contrast to the uncharitable
attitude towards the possessors of imperfectly declared or unde-
dared goods.25 The responsibility for declaring the goods rested
solely with the merchant. The publican hadno duty to be of assis-
tance with instructions or advice. Only, one readschillingly, "care
23 Dig. (UIpian) 39. 4. 1.4 : "et restituendi f'acultas publicano vi abreptum datur, quad
si t'ecerit, omni onere exuitur et poenali actione ex hac parte edicti liberatur. Dig.
39. 4-5-P''-. a preserved comment of Gaius, starts off by denying the right of restitution
after trial has been granted. Inthe next sentence, henevertheless ends up concluding the
opposite. This indicates that the text might have been interpolated. But when and how,
it"is difficult to say. It might just be an abbreviation of a longer argument eventually
leading to a less rigorous interpretation of the edict of the urban praetor. Tojudge from
Ulpian's very general statement, there is no compelling reason to take the opinion as a
post-classical invention.
24 On a similar note, the expectation of Cottier 2003: 226 that a decisionfrom AD 5 to
remove casesconcerningtheAsianportoriumfrom thegovernor'scourt to thepraetor
peregrinus in Rome would have enabled a more equitable administration of justice,
cannot be taken for granted. The Roman publicani would have been well connected in
the imperial capital. So that would not have prevented the working of influence. What
is certain, however, is that the threshold for seeking redress via the courts would have
been raised substantially. The merchant would now have to go all the way to Rome, a
much more time-consuming and costly enterprise. Taking on the publicans had become
more difficult, cf. Engelmann and Knibbe 1989: 120.
25 Seede Laet 1949: 437-444 for a basic treatment. It is symptomatic ofthe asymmetrical
legal position of publicansand merchantsthat de Laetcould makedo with less than
two pages for actions against publicans, whereas he spends seven pages paraphrasing
the titles in the Digest against smugglers. See Klingenberg 1977 for a more thorough,
butjuriciically over-systematic analysis.
209
 IMPERIAL BAZAAR
should be taken that he does not mislead those who are willing to
declare".26Falselydeclaredorundeclaredgoodswere,asa general
rule, simply confiscated. 27 If an error in declaration had resulted
in the tax collector getting more than his due, it was apparently
even necessary for the emperors Severus and Antoninus (Cara-
calla) to specify that the merchant should have the excess amount
returned. 28 There were other modest concessions granted in the
name of equity. Had bad weather forced a merchant to unload
his goods from a ship withoutdeclaringthem at customs first, he
was not to sufferconfiscation. Equally, a fragment of an Egyptian
customs law states, if the tax collector had ordered a cargo to be
unloadedwithout finding any goods not declared, he should carry
the costs of unloading himself. But such mitigating measures were
26 Dig. (Marcian) 39. 4. 16. 5-6. According to Tacitus (Ann. XIII.5I), Nero declared that
all customs laws should be made public so people could at least read them. To what
extent this happened it is difficult to say, but the very need to promulgate such a
reform speaks volumes. The inscription with the Asian Customs Law, dating from
AD 62 and found in Ephesos, shows that some inscriptions certainly were set up fol-
lowing the Neroniandecision. But, considering the great scarcity of finds of customs
laws from all over the Empire, it is doubtful whether the reform was generally imple-
merited. The Asian Customs Law was published by Engelmann and Knibbe 1989. A
revised version is available in Merola 2001, appendices 1-4. A new edition is now in
preparation.
27 Asian Customs Law (MonEph. ), lines 45-56 with Dreher 1997: 89-91. See also (Ps)-
Quint. Declam. 359: "Quadprofessusnon fuerit, perdat"and P. Oxy. 36. Klingenberg's
attempt (1977: 62-65) to introduce a distinction between non-declaration and false
declaration where only the former would lead to confiscation can now safely be rejected
on the basis of the Asian Customs Law, pace Dreher 1997: 89, n.45 who attempts to
rescue the interpretation by seeing it as a later legal development. There is no good
reason to take such a view. Klingenberg's interpretation was based on a priori reasoning
with no basis in the legal evidence. In fact, his own discussionforced him to recognise
that a falsely declared quantity fell under the right ofconfiscation. That left only the false
declaration of quality. In that case, Klingenberg maintained, declaration had actually
been made and the goods would therefore not have been subject to confiscation. I
doubt Romancustoms officials wouldhavebeenmuch impressedby such an argument.
A merchant attempting to pass off some bales of precious silk as a few bundles of
coarse, inexpensivecloth would surely, on detection, have seenhis goods confiscated
for smuggling. The Asian Customs Law confirms this. The act of making a declaration
(professio/apographe) is treated as requiring a statement of kind, value, weight and
number (Asian Customs Law (MonEph. ), § 18, and line 50 in particular). The problem
raised by Klingenberg, however, is a real one: when could the obligation to declare be
held to havebeenmet? Some legitimate disagreementamongthe parties over estimates
of value, for instance, must have been allowed for. It is symptomatic of the favourable
position accorded by the law to the publicans that the legal material practically left open
the question of when the duty to declare had been fulfilled.
28 Dig. (Marcian) 39.4. 16. 14.
210
A THIN LINE: PORTORIUM, PROTECTION AND PREDATION
few andyielded only cautiously. 29 Favour remained firmly on the
side of the customs collector. In a case concerning false declara-
tion of some slaves, the emperors Marcus and Verus "gracefully"
conceded that the poor wretch could get away with paying double
customs, it the failure to declare was unintentional, by mistake
rather than fraud. 30 Though certainly far preferable to outright
confiscation, this was still a harsh penalty. Double duty was the
sanctionimposedonpeoplewhofailedto defraythecustoms after
declaration had been made. 31 But in this case, it was to be paid
on some slaves who would have been exempt had the "error" not
occurred. As far as one can tell from the context, the slaves were
for personal use and could thus normally have been passed free
ofcharge.32 Confiscatedgoodswere, as a rule, auctionedoffwith
the profits accruing to the tax collectors and the authorities. The
previous owner might even be offered the opportunity to buy back
his lost property. It does not require much imagination to see how
theright ofconfiscation andthepractice ofselling backthegoods
to the original owners might sometimes have been turned into a
profitable system of abusewherepeople wereforced to "ransom"
their own property. 33
The wronged party could of course always attempt to get his
case heard in the court of the often distant Roman provincial
authorities. This did undoubtedly happen, as the legal responsa
preserved inthe Digesttestify. Butlitigationwasandstill is a slow
and burdensome way of resolving conflicts. In addition it could
29 Dig.(Marcian) 39. 4. 16. 8;P. Oxy.36.Klingenberg 1977:70-82makesthemostofthese
few concessions.
30 Dig.(Marcian)39.4. 16. 10.DeLaet1949:443,followedtosomeextentbyVittinghof
1953:395-397,isnotjustifiedinmakingthisa generalprinciple.Bothscholarsignore
thecharacterofthejudicialpronouncementsin theDigest.Thepronouncements"often
representjudgements inconcrete casesandfarfrom alwaysadduptoa coherent system
without internal contradictions.
31 Asian Customs Law(MonEph.), § 38.
-12 Dig. (Marcian)39.4 16. 10follows in closeconnectionwith39.4. 16.9 concerningan
error made by a minor in declaring the slaves for his personal use. The slaves to
be returned againstdoubleduty are thus best seento belongto the samecategory.
Cf.CodexJustinianusIV.6I.3 exemptingsoldiersfromsufferingconfiscationofduty-
exempt goods. In general, however, goods exempt from duty stifl risked confiscation'if
wrongly declared, see Kiingenberg 1977: 70.
33 Dig. (Paulus) 39.4. 11.4, cf. 394. l6. pr. Klingenberg 1977: 115-116treats this as an
issueof whetherthe dominushada rightto buybackhisproperty. This is besidethe
point.
211
 IMPERIAL BAZAAR
be fairly risky to complain about an abusivetax collector. First of
all, it was far from unlikely that the provincial authorities were in
collusion with the publicans. 34 Second, many customs officials
would have had a relatively sizeable staff of slaves and other ser-
vants at their disposal. These came in handy during situations
of conflict. T\\efamilia publicanorum constituted a ready instru-
mentofpowerwhichmightexpedientlyconvincedissatisfied"cus-
tomers" to hold their tongue, pay up and move on. One guard, for
instance,reportingaboutfraudulentcustomscollectioninanEgyp-
tian village, wasdelivereda severebeatingby the local publicans
in return for his efforts. A central concern in the section of the
Digest dealing with publicans is how to treat abusive conduct of
their slaves and other dependants. 35 Not all publicans, of course,
need to be seen as equally powerful. There is a greatdistance sep-
aratingthe big and notorious Roman companiesfarming some of
the imperial customs across large districts and the small contrac-
tors collecting dues at the gates of Egyptian villages. But whether
big or small, thepublicanshadto besufficientlywealthyto pledge
property to the state in guaranteeofthecontract. They wouldhave
had to be men of at least some substance in their communities,
though far from always members of the political elite. 36 Thus,
whether a specific incident was ever reported to the courts might
come downto a questionofthepowerandinfluencepossessedby
thepersonsinvolved anda considerationofthecostandtrouble of
judicial proceedings set againstthe sum required to strike a deal
withthepublican.Thebalancemustoftenhavecomeoutin favour
of bending and putting up with the publicans.
34 Cicero's Verrine orations are essentially a catalogue of the various ways this could
happen. The phenomenon has been dealt with in Chapter 2 above.
35 P. Amp. 77 = Wilcken, Chr. 277 = Johnston 1936, no. 348 for the attack on the
guard reporting irregularities. Dig. 39.4. 1.5-39.4. 3 and 39.4. 12 deal with thefamilia
publicanorum.
36 LexIrnitana(Gonzales 1986),chapters63-64forbidsmembersofthelocal council (as
well as their close kin and dependants) to contract for local vectigalia. This practice,
however, was not observed empire-wide. See Andreau [974 for a study of a man of
middling wealth contracting for some vectigalia in Pompeii. On the other hand, rich
councillors might on occasion bid for the collection of vectigalia outside their home
town. A member of Antioch's council is known to have contracted for the collection of
theimperial customson EasternluxuriespassingthroughPalmyra, Inv. Palm. X, 29(R.
Drexhage 1988, no. 29).
212
A THIN LINE: PORTORIUM, PROTECTION AND PREDATION
Parasitical fiscalism
If legal regulations were heavily tilted in favour of the customs
officials,it wasbecausetheinterestsoftheimperialstatehadmore
in common with those of the tax collectors than of the merchants.
Customs duties were not, as they later became in mercantilist
strategies, an important tool whichthe government usedto shape
and promote economic activity in the empire. The main interest
of the state in the exaction of customs was fiscal. 37 Trade was
to a large extent viewed as a flow of resources which could be
tapped into for the sake of revenue. Hence, in order to maximise
their proceeds, neither the Romans nor the Mughals did much
which would lessen the appeal of handling the levy of duties.
This has made the two tributary empires look very strange indeed
to modern commentators. Customs were not something mainly
charged on exports, and especially imports, in order to work as
an economic barrier between a state and the outside world. The
imperial systems taxedtheinternal movement ofgoodsas well as
ofthosecrossing theirexternal boundaries almost indiscriminately.
The empires did not totally abstain from graduatingthe customs
and excise rates. But it happened for very different reasons and
along quite different lines. In general terms, provisioning policies
and consumer interests were more prominent in shaping policies.
Tributary elites put greater store on well-suppliedmarkets, to the
equaladvantageofurbanconsumptionandtaxationpotential,than
on guardingthe interests of any specificgroup ofmerchants.
The sharpest illustration of this is the concessions granted by
the Great Mughals to some of the European chartered compa-
nies trading on India in the seventeenth century. These enabled
European traders to participate more or less on equal terms with
the native population in commercial pursuits. Indeed, from a
mercantilist point of view, the privileges had the absurd result
that in some periods the European merchants even seem to have
enjoyed some advantages incomparison to theirIndiancolleagues,
37 Originally propounded by Cagnat 1882, the view has largely been confirmed by later
research such asdeLaet 1949 (with some exeptions); Vittinghof 1953; and France 1994
and 2001.
213
 IMPERIAL BAZAAR
particularly with respect to road tolls and sales taxes. 3 The lack
of external customs barriers is a slightly more contentious issue in
the Roman case. Though convincingly criticised when it was first
aired, Sigfried de Laet's hypothesis of the creation of an external
customs barrier, especially on the eastern frontier of the Empire,
has refused to die out and is occasionally cited with some approval
in scholarly discussions. 39 The basis of this theory is the imperial
levy of 25% ad valorem charged on the import of Eastern luxuries
in the provinces of Egypt, Syria and Arabia. There is, however,
not one scrapofevidencein support ofthe view thatthis very high
duty was imposed to impede commercial exchanges with the East.
Some three decades ago Paul Veyne demonstrated the absence of
a mercantilistideology in the Empire.40Thereis, too, the unequiv-
ocal testimony of Strabo, the geographer. In his portrait of Egypt
the high duty on Eastern luxuries is treated as part of a discus-
sion of the large proceeds generated by the Roman administration
in the province. The rich trade with the East is singled out by
Strabo for its revenue-generating qualities: "as far as India and
the extremities of Aethiopia, from which the most valuable car-
goes are broughtto Egypt, andthence sent forth againto the other
regions; so that double duties are collected, on both imports and
exports; and on goods that cost heavily the duty is also heavy. "41
If customs were exacted at an onerous rate on Eastern luxuries,
it was because the trade could support them. The imperial state,
motivated by fiscal concerns, took advantage of this condition
and squeezed the merchants all the more tightly to exploit to the
38 Prakash 1998: 119-134 and Hasan 2004: ll3-l(9and [24-125 for the different sales
tax, road toll and customs rates in seventeenth-century Mughal India. More generally,
Pearson 1991: 96-98 and I 10-113 and Richards l993a: 203.
39 Especially France 1994: 142. See Dreher 1997: 85 for a more reserved endorsement.
For the original critique of de Laet, see Etienne [952 and Vittinghof 1953: 378-
384.
40 Veyne 1979.
41 Strabo XVII,line 13 (my emphasis). There are other examples of increasedcustoms to
take advantage of an unusually rich trade. One such is the attempt to tap into the wine
trade with Gaul in the last century of the Republic, reported by Cicero in Pro Flacco
9, 19-20 ("pecuniam permagnam ratione ista cogi potuisse confiteor", my emphasis).
The highest rate was charged on wine leaving the Empire as a luxury intended for
barbaricum. Yet another instance is the taxing of purple shells at twice the normal rate
in the Asian Customs Law (MonEph., line 20).
214
A THIN LINE: PORTORIUM, PROTECTION AND PREDAT10N
full the revenue-generating potential of this particular branch of
trade.42
In the absence ofa strong, systematic mercantilist impulse in the
tributary empires, the fiscally dominated approach to the exaction
of customs did nonetheless on occasion give way to other objec-
tives. 43 In both empires we find examples of the state extending
customs privileges to some groups or activities. 44 But concessions
wererarely, ifever,grantedwitha strongwishtoremouldtheeco-
nomic system. The motives were different, mainly political or ide-
ological innature. IntheMughal Empirethisisbroughtoutclearly
by the attempt ofthe last great Mughal emperor, Aurangzeb ( 1658-
1707) to favour Muslims over Hindus withregard to the payment
of imperial excise duty, the wkat. In afarman of 1665 he set the
rate for Hindus at 2/40 and for Muslims at 1/40. Two years later
he went even further and abandoned the duty for Muslims alto-
gether. Thishappenedaspartofa policy to boostMuslim imperial
primacy. The policy, however, had unexpected consequences. In
the eyes of the emperor Muslim merchants abused their newly
acquired privilege. They beganto handle the goods of their Hindu
colleagues on payment of a commission fee. That way the Hindus
avoided the imperial excise duties and the Muslims made an extra
profit. The loser was the Mughal treasury, whichsaw the income
from customs dwindle. This was not what the government had in
mind. The exemption was intended as a political and ideological
symbol of the religious supremacy of Islam - "that the Muslims
may be distinguished with obligation and kindness from vicious
unbelievers", as an imperial order stated. The privilege had not
been granted with a view to creating a Muslim commercial dom-
inance in the Empire. Hence in 1681, after a little more than a
42 For an analysis, looking at archaeological traces of customs exactionon the eastern
trontjer, also emphasising the predominance offiscal motives, see Young 2001: 66-74
andthegeneraldiscussionin chapters5 and6.
43 0rsted 1985:288-289triestodetecta protectionist dimension inthefiscallymotivated
exaction of customs. This is unconvincing and is based purely on speculation. All
exactionof customstendsto makethemovementofgoodsmoreexpensive.Thisis an
inevitableeffectoftolls andcannotbeseenin itselfastheresultofprotectionism.
44 France 1994:143-144isrighttoinsistthatRomancustomscannotbeentirelyreducedto
a fiscalquestion. Butadmitting thisdoesnotcommit onetogosearchingformercantilist
measures. Political and ideological objectives should not too readily be brushed aside.
215
 IMPERIAL BAZAAR
decade,the exemption wascancelledandimperial exciseduty was
reintroduced for Muslims.45
The same concern with limiting the fiscal effects of politically
motivated concessions occurs frequently in Roman policies. The
Asian Customs Law recently found on an inscription in Ephesos
contains several examples. Among the privileges granted to the
colony of Augusta Troas, (re)founded by Augustus, was the grant
ofthe rightto the imperial customs on goodsimported or exported
from its territory. However, goods which were passing through in
transit, still fell underthe rights of the publicans appointed by the
Roman state.46 The aim, in other words, wasto grant an economic
privilegeto a politically favouredcommunity, notto makeAugusta
Troas a dominant emporium. The same goes for the exemption
from customs for thirty days granted on imports, but not exports,
to the territory of most probably Pergamon and its harbour city
Elaea, in order to facilitate supplies for the politically important
festival of the Romaia Sebasta.47 If we change our perspective
from individual grants of privileges to general imperial policies,
the impression is confirmed.
The Roman state did exempt some groups and activities from
duty across the Empire. This comprised, among other things,
all transports of state-owned "goods", and goods belonging to
45 Mir'at-i Ahmadicontains the Mughal farmans dealing with the gradation in the rate
of duties. See chapter 121 and the following for the cancellation of Muslim customs
exemption (citation in chapter 121); chapters 104, [o8 and more broadly 97 and 120
for the introduction of Muslim privilege, cf. Hasan 2004: I 17. For a broader treatment
of Muslim religious privilege, see chapters 97 and 120 of Mir'at-i Ahmadi; Richards
[993a: 171-177;andfor some modem treatments, see Hodgson 1974, vol. 3:92-98.
4 Asian Customs Law (MonEph. ), lines 103-105 with Engelmann and Knibbe 1989:
114-115.
47 AsianCustoms Law(MonEph. ), lines 128-133withanalysisin EngelmannandKnibbe
1989:125-129. Seede Ligt 1993:42-48 andparticularly 222-234for the dominanceof
symbolical and consumerist motives involved in the granting of festive ateleia combined
with a strong continued imperial fiscal interest. This. of course, does not rule out that
such privileges may also occasionally have turned out to be of benefit to transit trade
as well. But it was not normally the main purpose. The elevation of Delos, by the
Roman authorities in the second century BC, to the status of a free harbour to harm
the central position of Rhodes in the eastern Mediterranean trade may be thought to
constitute evidence to the contrary. But not even here was trade the motivating force.
The action was taken as punishment for wavering Rhodian support during the Third
Macedonian War and included stripping the island state of her Anatolian dependencies
as well, Scullard 1980: 287.
2l6
A THIN LINE: PORTORIUM, PROTECTION AND PREDATION
military personnel and veterans. 48 Such exemptions were used, no
doubt to some extent successfully, as we would expect from our
Mughal parallel, to gain commercial advantages. 49 Private con-
tractors transporting state goods, for instance, attempted to mix
their own goods with the official cargoes to avoid paying customs.
Soldiers would engage in trade andexploit the exemption to gain
the edge in the competition with private merchants. These were
inevitable consequence of political privilege. In the view of the
imperial state, they were also unfortunate, causing an unintended
lossofrevenue.Repeatedlythestatetriedtocontainthesepractices
anddiminish its economic loss. 50The emperor Hadrian instructed
provincial governors, legionary commanders andtheirprocurators
when they dispatched people to buy things for their personal use
to write a memorandum for the customs collectors indicating the
exact amount ordered "so that anything brought in excess can be
subject to duty" 5' in the Asian Customs Law we read generally
that the exemption for state transports also included the people
who operated them. But only on items for their own use. This was
similar to the arrangement for soldiers. Their freedom from excise
was limited to items for personal use. If the goods were intended
for trading, customs were due according to normal rules. 52 When
thetributary stategrantedcustoms privileges it wasnot inpursuit
of a general economic policy. The goal was more limited. ' Privi-
lege wasawarded ina more narrow trade-off thatjealously sought
48
?^. altl^ 427-^3'and 432-435 for a broad survey of the material.
Normally the
instruments oftravel wereexemptfor all persons.
49 As pointed outby Whittaker 19940, chapter4.
50
Inaddition
^to the fo llowing examples; Dig. 39.4.9. 8, excerpted from the spurious
entiw Pauli,^ deserves mentioning. It states that while the imperial fiscus'is free
from all duties, the merchants who market imperial goods cannotclaim any general
exemption from customs.
51 Dig. (Gaius) 39.4.5.i.
2 AsmncustomsLawWonEph. ),lines58-66.Tac.Ann.XIII,51.Theprivilegeofsoldiers
illater.me"tione(l'"codexJustinia'wsIv 61,7 (discussinga lateimperial~'sales"tax)
andagaintheyarenotexemptinmattersofbuyingandselling.Constantine'intr'oduced
immunity for all navicularii. the transporters serving the state, even for goods
;"te"dedto sel1(c n-XI11- 5'5 with de Laet I949-:480 andDe Salvo ^2:259).
Butthetitle ofnavicularius wasconnected withobligations andit issymptomaticthat
theRoman state,just as the Mughal, did not look favourably uponthe'habitof some
naviculariitocarrythegoodsofothers.HenceC.Th.XIII,5,'24'specifies:"naviculam
areexempt whenit isproved thatthey dobusiness fortheirownaccount (cum sibi rem
gerere probabuntur)".
217
 IMPERIAL BAZAAR
to safeguard the state fiscal interests while giving recognition to
groups and individuals that for onereason or another were deemed
importanttothepoliticalorder,beit, forexample,assoldiers,gov-
emmental transporters or as those involved in significant religious
activities.
The imperial state did not havemuch incentive for changing the
predominantly fiscal approach to the commercial sector. Neither
in the Romannor in the MughalEmpirehadexpansioneverbeen
closely tied to collaboration with commercial groups. Therefore
thetributary statehadnocompelling reasonstoforgorevenue inthe
short-term by favouring specific groups of traders. As the empires
grew and the amount of trade inside the boundaries expanded to
ever greater dimensions, compared to exchange with the world
on the outside, this became even more the case. From the per-
spective of the treasury there was not much to be gained, except
loss of revenue, from helping one group of merchants to capture
a greater share of a largely internal commerce at the expense of
other traders in the empire. 53 Instead, the exaction of customs was
directed to take profit on existingtrade flows irrespective of their
direction. Hence excise and other trade-related dues were regu-
larly levied in nodal points or on traffic arteries where many goods
more or less had to pass, such as at city gates, harbours, high-
ways,river crossings,bridgesandin thebazaars.54Customs were
in the Roman world to a very large extent passagetolls. 55 It was
not uncommon to levy the tax on the same goods both on entry
53 The first European mercantilist "economists" saw the global economy as a zero sum
game too (Anderson 1974). However, as the states were much smaller the advantage
to be gained from capturing a greater share of the international economy at the loss of
foreign competitors was all the more apparent. Only later, particularly in the eighteenth
century, did the notion of real growth develop on the basis of these ideas Though
it would be wrong to claim any automatic link between full-blown mercantilism and
small states in a partly internationalisedeconomy, it still seemsclearthat commercial
interests could bebetter heard in such systems. Thus without developing real mercantilist
policies,theMughalsuccessorstatesineighteenth-centuryIndiadidgivemoreattention
to commercialgroupsthantheir imperial predecessor(Bayly 1988: 1-78).
54 P. Oxy. 1439-1440;ILS375 (collection at city gates); AsianCustoms Law(MonEph.)
chapter 21 (customs office close to the landing of ships); P. Oxy, 1650 (Memphis
harbour tax, excise at administrative district border); OGIS 674 (road tax); AE 1975:
413(riverbank); Dig. 19.2.60.8 (bridgecrossing);P. Land.Inv. 1562versa(markettax).
55 Now again confirmed by France 2001, summarised in the conclusion. However, France
seems to me to underestimate the possibility of double customs payment, cf. note 57
below.
2l8
A THIN LINE: PORTOR/I./M, PROTECTION AND PREDATION
and exit through a customs district or station. 56 Consignments of
goods travelling beyond the confines of their nearest region would
also normally have had to pay several tolls, sometimes at a consid-
erable number of stations. Customs multiplied with the distance
travelled. -57The tendency for excise duties to accumulate is aptly
illustrated by some tax receipts found among the Egyptian papyri.
56 DeLaet 1949andVittinghof 1953:375-378, andnowconfirmed bytbeAsian Customs
Law, which applied to both imports andexports, cf. Dreher 1997. However, the notion is
difficult to reconcile with modem sentiments and doubts continue to be voiced without
any good reasons, e. g. France 1994: 138-139. A letter from Cicero (Alt. II, 16, 4)
reveals thattheAsianpublicanieventriedtomakemerchants paytheAsianimport duty
againwhenthey moved from one city to anotherinsidethe province. It may also be
notedthatthe AsianCustoms Law(MonEph., lines [6-20) exempts people from paying
imports or exports twice on the same goods, in the same year, at the same customs
station. Thus, if a merchant had taken his goods to a city in the province of Asia and
paid the import duty, if he then left againwith some of his goodsandreturned later in
the yearhe would not have to pay customs on the same items again.Thepublishers
(Engelmann and Knibbe) typically prefer to take this to mean that the merchant was
exempted from paying re-export duty on goods already paid for at entry. This is not
supported by the inscription which clearly deals with exports and imports separately.
On Egypt, see also Sijpesteijn 1987, chapter 5, who accepts the principle but believes
some goodsto be exempt, especiallywine, whichhetakesto betaxedonly on entry.
But a look at Sijpesteijn's data makes it clear that his conclusion is more likely to be
a reflection of the prevailing flow of goods in the villages whose records have been
preservedthanevidenceof avoidanceofdoublepayment.Thuswinewasimportedin
small quantities, whereasoil was sentout. Hence wine is documentedin relation with
imports andoil withexportsfrom the villages.
57 Rouge 1966: 450, emphasising the accumutative burden of excise dues. See further
0rsted 1985: 285-289, followed by De Salvo 1992: 317-321 and France 1994: 140.
Drexhage 1994 and Drexhage, Konen and Ruffing 2002: 145-147 underestimate the
number of customs exacted. They strangely ignore the evidence from Egypt, where
customs receipts reveal the district duty of the Memphis harbour tax being added on
topot city customs(cf. n 59). Purcell2005insists,tomy mindrightly, ontheubiquity
of local customs, yet maintains they did not accumulate (pp. 220-222). That seems
counter-mtuitive not to say a contradiction in terms. The customs inscription from
Caunos (SEG XIV: 639), cited as evidence, does not prove that customs were never
paid on goods in transit. Quite the reverse, the inscription (very fragmentary) describes
a regime of exemptions paid for by a local benefactor, propably as a suspensionto
encourage imports in connection with a periodically recurring religious festival (cf.
Schwarz200ib: 288-289). Othercustoms laws, such astheregulations concerning the
portorium collected in Asia, do not as a rule mention a specific exemption of transit
goods. Indeed both the customs regulations preserved from the Lycian koinon and
the province ofAsiaenvisage taxing transit goods. Augusta Troas andMyra eachhad
the right to collect these respective customs on goods sold within their own districts.
Thosejustpassingthrough, however, hadtodefraydutytothecollectors oftheprovince
andthekoinon(MonEph.,lines 103-105;Schwartz200ia). Equally, Purcell refersto
the transit dues collected by Massilia at the mouth of the Rhone (Strabo 4. 1. 8). But
merchantsalsohadtopaytheimperialcustomsonenteringtheGallicprovinces.Dues
would have been accumulating with distance.
219
 T
IMPERIAL BAZAAR
By chance, several receipts have survived documenting the pay-
ment of different customs on the same consignments of goods.
Thus we find small amounts of wheat, travelling a fairly short
distancenorthwardsfrom the village ofPhiladelphia,payingthree
separateduties: onefor leaving the nome, onefor usingthedesert
road and one for the so-called Memphis harbour tax. 58
It is noteworthy,too, howthe different Egyptiantolls tendto be
connected with different levels in the state system. The Memphis
harbour tax was charged on goods crossing between the larger
administrative districts, the so-called epistrategia, while there was
also a more local tax for leaving the primary administrative unit,
the name. This is a typical feature of segmentary state systems.
They consisted of many parts which both tried and had to be
allowed to tap into the economy and get their share of the tribute.
An analysis of the structure of Mughal customs collection in the
province of Gujarat makes the principle clearer. The central state
exacted a customs-cum-sales tax on all movement of goods in the
province. In addition to this there were numerous local tolls of
various kinds. These were controlled by the various aristocratic
groups thattogether madeup the state on the ground. The central
Mughalstatedidformallyclaimtoexerciseauthorityoverthelocal
levies. Some of them were even gradually being drawn into the
orbit of the imperial system, as they became part of the revenue
assignments of Mughal nobles, the so-called jagirdars. But that
was as far as it got. When the state attempted to exercise closer
control, for instance by outlawing local dues or reducing their
number, it encountered stiff opposition. The Mir'at-i Ahmadi -
a history of Gujarat - reports how on one occasion aristocrats
demanded economic compensation if they were to accept such
an abolition. In that situation there were essentially two options
open to the state to defend its policy, neither of which was very
attractive. It might try to hold its own and compel the noblemen to
agree to respect the ban. In the long-term this was bound to be
ineffectual. Aristocrats were not generally inclined to accepthav-
ing their incomes reduced, especially not when they were running
58 P. Customs, list l, nos. 139-141,seealsonos. 184-186, 355-356, 362-363and4l9-42l
with Sijpesteijn 1987: 19.
220
A THIN LINE: PORTORIUM, PROTECTION AND PREDATION
things on the ground. Alternatively the state might grant the aristo-
crats compensatory incomes. As a general solution, however, this
was equally problematic in the eyes of the imperial government;
it was expensive and diverted scarce resources from other activ-
ities. So either way, the government tended to back away from
the abolition of local dues and allow their continued collection
or silently tolerate their reintroduction after a short term of sup-
pression. The revenue pressure exerted by the different segments
of the Mughal polity consolidated the existence of a customs and
excise regime with frequent collection points and the co-existence
of many separate dues. 59
This mechanism can be seen at play in the Roman Empire, too.
During the high Empire most regions of the realm were divided
into a number of large, separate customs districts where toll was
collectedinthenameoftheRomanstateonentryorexit.Theexact
nature of these districts is not entirely clear, but they bore some
relation to provincial divisions, without necessarily being strictly
identical. The Asian portorlum, for instance, was clearly arranged
aroundthe core of the old Attalid kingdomthat madeup the bulk
of the province of Asia Minor. But to this were added, among
other things, the customs collected in the strait of the Bosporus,
even though this area was outside the provincial boundaries of
Asia. 60 Beneath the umbrella of imperial excise duties the exis-
tenceofa systemoflocallybasedtolls andlevies is attestedby our
59 Mir'at-i A. hmadi, chapter 117. See Hasan 2004, chapter 7, for the customs system in
Gujarat. Pp. 114-12i areparticularly relevant forthe analysis presented here. In general
about the many tolls in the Mughal Empire, see Habib andRaychauduri 1982: 186-189.
TheEastIndiaCompanyaftertheconquestofBengalattemptedthecancellationofsuch
tolls but ultimately foundtheresistancefromlocal elitegroupstoo strong (Chatterjee
1996, chapter 5).
60 AsianCustoms Law(MonEph. ), lines 22-28 defines the Asiancustoms district; seelines
8-ll and 13-15for exactionofcustoms in thestraitoftheBosporus, cf. Nicolet 1993:
945-951- Some customs districts seem to have comprised several provinces, though
without losing all relation to the administrative framework, as in the cases of Gaul and
lllyricum, cf. de Laet 1949 for these larger circumscriptions. France 2001 treats the
Gallicportoriumanew.He rejects the notionofcircumscriptionandconcludesthat the
qiiadmgesima Galliarum was a passage toll separating Italy from the north-western
provinces. I am not entirely convinced. He seems to be misled by the chance preser-
vation ofevidencein the Alpine regions. See0rsted 1985:251-347 for an interesting,
but speculative and over-complex analysis of the material pertaining to the Illyrian
portonum.
221
 IMPERIAL BAZAAR
sources. To someextentthishasmilitatedagainstthepreconcep-
tions of modern researchers expecting to find a rigorously defined
legal, coherent administrative system in the Empire. They have
therefore suggested various ways around this so-called problem.
For instance, it has been claimed that these tolls were primar-
ily collected by communities that in legal terms were outside the
imperial system, the so-called civitates liberae ac foederatae.
There is no strong evidence in support of that proposition. One of
the best-preserved inscriptions with an example of a local customs
law was found in Palmyra. The inscription was erected in the reign
of Hadrian. At that time the city had formally been part of the
province of Syriafor more thana century.63
Another way out of the "predicament" has been to claim that
local tolls were a feature of the Greek east, whereas the Latin west
waslargely withoutthis phenomenon.64Thatis unconvincing.An
octroi wascollected at the very gates of Rome, the capital. 5 It is
undeniable that in terms of sheer wealth of epigraphic testimony
local dues are far better attested in the eastern provinces than else-
where, but this is more likely to be a function of the epigraphic
record than anything else. It is worth reminding ourselves that the
sources are fragmentary traces rather than a direct reflection of
France 1999 admirably updates the material first gatheredby de Laet 1949: 351-361.
See Schwarz 200 ib, III, 4 for an excellent treatment of local customs exaction in
Anatolia. Outside the province of Egypt, the large rabbinic literature connected with
Judaea has yielded some of the best evidence for the density of customs collection, cf.
Goldschmid 1897: 199-202.
62 France 1999: 99-100 (and more broadly France 2001) surprisingly reverts, it seems, to
the notion of Rostovtzeff 1902: 390 (and, it would seem Frank 1933-40, vol. [: 255)
that the right to collect local portorialvectigalia was restricted to civitates liberae ac
foederalae. This view had already been exploded by de Laet 1949: 90 and 351 (and
now confirmed by Schwarz 200 ib: 338-406), who observed that formal legal status was
not decisive. Quite the reverse, we even find attempts to levy the imperialportorium in
"free" communities, e. g. Byzantium.
3 Young 2001: 137 summarises the evidence on the status of'Palmyra: "the city should
accordingly be understood as a tributary city of the province of Syria". Within modern
Palmyrene studies, the notion has gainedcurrency that the city became a civitas libera
following a visit ofHadrian, during which the city changed its name to Hadriana Tadmor.
This doesnot affectour argument sincethe tax law, as is clear from the inscription, also
existed during the first century of Roman rule. However, as observed by Fergus Millar
(l993: 325)> this modem assenion is completely baseless and may safely be discarded.
Under Septimius Severus, the city was made a Roman colonia.
4 DeLaet 1949: 353 andFrance 2001: 278-283, whoessentially restricts thephenomenon
for all Gaul to Massilia.
65 CILVI, ioi6a = ILS375 with de Laet 1949: 347-349.
222
A THIN LINE: PORTORIUM, PROTECTION AND PREDAT10N
the ancient reality. The challenge is to interpret the nature of the
lacunae in the material. Ifpositive evidence is scarce in the western
Empire, there are nevertheless indications that local excise duties
were asmuch a part oflife there asthey were in the Greek-speaking
east. Cities across the Empire were expected to possess their own
vectigalia - taxes accruing, for example, from the leasing ofcom-
munal lands, pasturageand,not least, variousdueson commercial
traffic. It seems improbable that cities in the western provinces
should, en bloc as it were, have abstainedfrom levies of different
kinds on the movement of goods.66
Imperial conquest, in other words, did not eradicate the exis-
tence of locally administered dues on trade. The conquerors toler-
ated their continuance. As inthe Mughal Empire, local government
needed financing and the provincial aristocracies did not bear the
entire burden themselves. Formally, the Roman state did claim
some sort of sovereignty over these tolls. At the time of inclu-
sion in the imperial fabric, the local community would normally
havereceivedtherightto continueorbeginthecollection of some
such taxes. In the eyes of the Roman state the post-conquest set-
tlement could not be changed without its express permission. The
introduction of new duesrequired imperial sanctionbasedon the
recommendation of the provincial governor, as Vespasian made
clear in reply to a request from the Baetican community Sab-
ora. 67 Gradually the "hunger" for tribute led the Roman state to
confiscate and absorb more and more ofthese tolls into the impe-
rial revenues. This happened slowly and piecemeal, apart from
in Egypt where the state had already successfully come to con-
trol local dues before the Roman annexation. 68 In the rest of the
Empire, the development culminated in the fourth century when
different emperors made several attempts to clear the table and
includeall localvectigaliain the imperialtaxes.69
66 ?orvecti8alia: .seeLexIrnitana(Gonzales 1986),chapter76;Sabora(CILII, 1423).
SeeRouge 1966:449for a similarpointofview.
^IIlI.4_2^', co"fiTled, by Dis' (Hermogenian) 394. pr
lo.
Cf.someofthescatteredcommentsindifferenthistoriansabouttheusurpationoflocal
vectigalia by emperors. Forexample, Suet. Tib. 49. 2; Vesp. 16.
69 Jones 1964:732-733,whodatesthe generalconfiscationoflocalvectigaliato the
reignsofConstantineandConstantiusII.Afteritsinitiation,thefortunesofthispolicy
fluctuated considerably. Julian abandoned it. Then it was reinstated. Then the cities
223
 IMPERIAL BAZAAR
That was far too ambitious. Even when the imperial fiscus had
achievedsome sort of control, it was still forced to recognise local
claimsonthesetaxes,bothin EgyptandtherestoftheEmpire.This
normally happened in the form of allocations to local institutions
ofa proportion oftheproceedsfrom a giventax.70Julian,in a tem-
porary setback for the expansive policy, went further and simply
restored the old vectigalia to the communities. 7' His successors
reversedthis decision.But it is doubtful if they everachievedmore
than partial success. Some local dues continued to evade imperial
authority as they had in the preceding centuries. From the first
through to the fourth century AD the emperor and his administra-
tion repeatedlyfoundit necessaryto enjoin local councilsthatthey
were not allowed to institute new customs and excise dues of their
own accord without imperial sanction.72 For the most part, impe-
rial control, when it worked at all, is likely to have been relatively
lax and discontinuous. The customs law found in Palmyra offers
a fascinating glimpse into this. The surviving version, inscribed
on stone in AD 137, set out to revise the regulations for the local
collection of tolls. However, it also preserved the older version
and a few rulings by Roman officials in former disputes. What
is particularly interesting to note is that a Roman governor had
apparently been present when the old version was agreed upon.
The new law, on the other hand, appears to be entirely the work
of the local council acting independently of the Roman provincial
authorities. 73 In the light of this, it is difficult to speak of a unified
customs policy in the Empire. Roman hegemony did not lead to the
creation ofa single customs regime. The central state only partially
were granted a proportion of the customs proceeds. Then these were again confiscated
and later granted yet again. I doubt if we should take this narrative as a reflection of
how things worked out on the ground. Rather the line of events suggests only partial
or limited implementation of the imperial enactments. For similar scepticism about the
success of these measures, see de Laet 1949:462.
70 For grants of part of the customs to the city councils, see C. Th. IV, 13, 5 and 7 and
Codex Justinianus IV, 61, 13. The market tax of the Serapeum in Oxyrhynchus (P.
Land. Inv. 1562 verso) wasquite possibly allocated to the temple (Rea 1982).
71 Amm. Marc. XXV,4, 15. A very important element in Julian'smeasurewasthe resti-
tution of temple property.
72 CIL II, 1423; Dig. 394. lo. pr; and Codex Justinianus IV, 62, 1-4.
73 Compare the decree of the council of Palmyra, which introduces the tax law, with the
headingof the old law given later in the text. In Matthews 1984 this would be section
la with 2a. See further Rouge 1966: 448 in a similar direction.
224
A THIN LINE; PORTORIUM, PROTECTION AND PREDATION
controlled the exaction of tolls; it had to recognise or silently tol-
erate a probably very significant number of additional local titles
to tax existing trade flows. If tributary empires did not generally
erect tail customs barriers betweentheir subjects, they did not do
much to promote long-distance trade by abolishing internal tolls
either. Instead,the fiscal interests of the different segments of the
imperial system were allowed to dominate and the movement of
goods was slowed by the need to defray the costs of a potentially
considerable number of duties.
Merchantresponseand bargainingpower
Yet, commercedidnotgrindto a halt. Thepublicanhadto balance
his short-term interest in squeezing the merchant against future
loss of revenue if the continuation of commerce wasjeopardised.
This gave the merchants a far from insignificant bargaining power
vis-a-vis thecustoms collector. The imperial statedidnot only find
it hard to exert a tight control over customs exactions; in addition
its representatives experienced similar difficulties in monitoring
and regulating trading activities. Merchants were not always easy
prey. Both they and their goods were mobile. This made them
difficult to tax.
Shouldthe imperialauthoritieshappento forget, thelessonwas
regularlyrepeated.InthelateseventeenthcenturytheMughalstate
attempted to change the procedure for the collection of a sales tax,
zakat.Whereasmerchantshadformerly paidthetax at thetime of
purchase, they were now ordered to pay at the time of sale. The
chroniclerof the Mir'at-iAhmadiexplains:
the collection of tax at places of sale was decided upon because the price of an
article is higher at that place than at that of purchase. It thus means an increase in
revenue.Butasthemerchantssoldmanycommoditiesatplaceswherecollection
of taxes was not regular, the government suffered losses. An order was, therefore,
issued a second time for its collection at places of purchase. 74
The problems encountered by the Mughal authorities in rais-
ing the sales tax mirror the experience of the Roman state in the
74 Mir'at-iAhmadi,chapter126(p. 284inthetranslationofLokhandwala).
225
 IMPERIAL BAZAAR
collection of the so-called collatio lustralis or chrysargyron, a tax
instituted by Constantine to be collected every five years from
merchants and other businessmen. The emperors, Theodosius and
Valentinianus prescribed "that no more should furtive commerce
render the merchant rare in the glorious cities and the throng of
traders hide in obscure and secluded places to the loss of public
revenue".75 The reach of public power, always a precarious thing
in the agrarianate world, was particularly weak outside cities and
the main traffic routes. The world of Mughal and Roman mer-
chants alike was not only one of traditional abuse, it was also one
of equally traditional freedom, to paraphrase the late Ashin Das
Gupta.7 Merchants, therefore, were often tempted simply to cir-
cumvent taxation by conducting business in some remote village
or bay far away from the toll station and hungry glances of cus-
toms officials. Quantification is out of the question, but smuggling
is bound to have been rife and to have gone largely undetected. 77
Relocating business to more remote areas did, of course, have
some drawbacks. The traders would lose the central-place services
offered by cities, such as a greater concentration of buyers and
sellers. They would also be outside the protection of the law
and hence more vulnerable to predation by bandits and power-
ful landowners. Smuggling carried risks, but it offered enough of
an alternative to be a serviceable defence against new taxes or
unusually abusive or frequent customs exactions. In practice, the
flexibility of the merchant did impose some restrictions on the
conduct of customs officials. It did not prevent arbitrary preda-
tion but in the aggregate it would have exercised a moderating
influence, both on the conduct of officials and on the total tax bur-
den; hencethe generally relatively low officialcustoms rates, most
lying between 2% and 5% in the Roman and Mughal Empires.7
75 C. Th., N. Val. 24: "ne ulterius furtiva negotiatio et clans urbibus rarum facial mer-
catorem et obscuris ac reconditis locis in damnum publicae functionis lateat turba
mercantum".
76 Das Gupta 1987: 136; Hasan 2004: 117-118 treats the episode in the Mir'at-i Ahmadi.
77 See also Faroqhi 1984: 54-55 for an instructive description ofthe difficulties in detecting
smuggling in Ottoman Anatolia, complemented by pp. 66-69 on the ready distribution
of stolen goods.
78 SeedeLaet 1949 on Roman customs rates. OnMughal customs rates, seee. g. Tavemier
1889, vol. I: 7; Mir'at-i Ahmadi, chapter 104 (between 1/40 and 2/40); Hasan 2004:
226
A THIN LINE: PORTORIUM, PROTECTION AND PREDATION
They were intended to tap into trade flows, not force them under-
ground or out of business. It was not only a question of smuggling.
The merchant could often choose between different itineraries to
bringhisgoodsto market.Themostrecklesspublicansorthemost
heavily and frequently taxed routes might find that they were being
passed by to the advantage of revenue takers on alternative, but
less expensive routes. 79
The relationship between merchant andcustoms official should
not be seen as one-sided.Ratherit wasbasedon a tenuousbalance
ofpower.A gameofcatandmousewasplayedoutwheretherules
were subject to constant renegotiation according to concrete cir-
cumstances. The result was the establishment of a fragile modus
vivendi. 80 Some trades and merchants possessed more leverage
thanothers. According to the situation, particularly attractive "cus-
tomers", as it were, sometimes managed to wring smallish or
moderateprivileges andconcessionsfrom the customs authorities.
Among such"fiscal"rebateswe findthepractice, in whichfamil-
iar merchants passing a customs station on a regular basis were in
somecasesallowedtopostponethepaymentoftoll till theirreturn
whenthey could settle the bill out ofthe profits derivedfrom sell-
ing their wares. 81 The large and valuable export of slaves from
Anatolia was taxed at a slightly lower rate than slave imports. 82
'I3-ll. 7''andRit:hards[993a: 175and203.Thesimilarityofrangeissurelymorethan
coincidental.
79 Theextremeexampleofthisphenomenonisofcoursewhenbanditsruthlesslydescend
on traders. In his Roman history (XIV, 2, 2-4) Ammianus Marcellinus describes how
Isaurianbrigands,after havingplundereda numberof merchantcraft, found that their
source of "revenue" dried up as ships simply avoided the coastal stretch under their
control. Mughal sources allow us to follow this in greater detail than Roman. From
Surat a letter survives from a governor trying to persuade traders, especially the Euro-
peancompanies, to return to Suratby promisingto stop overtaxing andprovide better
protection for caravans, cf. Hasan 2004: 35 and 50-51.
80 Sijpestein 1987:44, forinstance,documents howEgyptianpeasants/traders, knownto
the villagecustomscollector, wereableto pay theirtaxes in'advance.
Dig. (Marcian) 39. 4. 16 I 2 providesevidence ofthepractice ofpostponing payment. On
a similar arrangement for the English East India Company inSurat, see Hasan 2004:
116.
?'isc:ustomarytointerpretthehigherimportrate(AsianCustomsLaw(MonEph. ), § 41,
1. 98) as an attempt through customs to discourage imports of slaves to Asia (Drexhage
1994: 6; Dreher 1997: 85). However, the difference only amounts to a few sesterces
per slave. Withslavesoftenpricedat severalhundredHS, theextraduty cannothave
beenevenmoderately prohibitive. Therefore, I seethesmaller export duty asa courtesy
discountgrantedto a particularlyvaluableexport.
227
 IMPERIAL BAZAAR
In Palmyra a camel-load of unguents imported in alabaster ves-
sels was taxed at 25 denarii. But when the merchant left the town
again he was only charged 13 denarii. This was probably a kind of
rebate intended to encourage traders to bring the merchandise to
the city in the first place; unsold goods would be charged less than
the normal fare on departure.83 Luxury goods from India pass-
ing through Egypt were perhaps weighed with favourable scales
whenpaying someminor, local duesontheir wayto Alexandria. 84
In the same context, some merchants might secure the patronage
of persons of power inside the tributary system. This could pro-
vide such merchants with preferential treatment and perhaps even
sometimes, tax concessions. 85 That would be a plausible explana-
tion of the discount on exported slaves in the Asiancustoms law.
Merchants supplying the Roman elite with much wanted slaves
presumably usedtheirconnectionsto havethe concessionwritten
into the law. However, patronage rarely comes gratis. Bribes or
other reciprocating services would often have had to be paid as
a kind of compensatory tribute. The profits from such discounts
would thus have re-entered the tributary system to a considerable
extent. This is presumably what was taking place when a group
of ferrymen put up, probably in consecutive years, inscriptions
and statues honouring the customs collectors of Chios "for their
virtuous conduct towards them". The "honorific monuments", as
suggested by Onno van Nijf, "may have been the public face of
83 PalmvreneTaxLaw,CISII, 3, 3913, panel 3a, lines 19-31. AgainI amcertainthatthis
should not be seen as a customs barrier to imports. The practice of occasionally granting
a discountto goodsbeingre-exportedin orderto attractthemto themarketin thefirst
place is alsoknown from the inscription recording a local toll in Caunos in Lycia(SEG
XIV, 639 with the comments of Bean 1954). There a benefactor set up a foundation to
pay for a number of exemptions from the regular customs regime. See now Schwarz
20()lb: 385-394. Purpura 1985 unconvincingly argues that the Caunos prescriptions
were regular rather than exemptions from the generally established practice.
84 P. Vindob. G 40822 (second column). For analysis of the discount in the weighing
process seeCasson 1990.DeRomanis 1998isaninteresting ifultimately toospeculative
attempt to make something out of these discounts. Rathbone 2001 now challenges
this interpretation and substitutes a scenario with surcharges accruing to the customs
officials.
85 In the law codes it is often observed that merchants connected with great households
managedto escapetaxation (C. Th. XIII.1.5; 15 and 25). However,the phenomenon
should not be seen as restricted to late antiquity. Cicero (Fam. 13. 75) shows aristocratic
patronage securing privileges in dealing with the authorities for a grain transporter/
trader.
228
A THIN LINE: PORTORIUM, PROTECTION AND PREDATION
personal corruption".86 Generally, however, fiscal rebates were
relatively fragile. Based on the concrete balance of power they
were permanently under pressure, subject to renegotiation and, if
the Mughal evidence is anything to go by, constantly subject to
erosion. 87
The toll of the Fax Romana- drawingup a balancesheet
So far my discussion has focused on the parasitic dimension of
customs exaction. But the tributary system can, to some extent, be
seen as offering the merchants something in return for the payment
of dues: a modicum of social order, protection from brigandage
and piracy. The Roman emperor claimed to have brought Fax to
the subject populations of the Empire; in the process roads and
seas had allegedly been cleared of pirates and bandits and made
safe for merchants and travellers. The imperial propaganda of the
Fax Romana has been accepted by most generations of classical
scholarship. The notion of an imperial peace has also occasionally
found its way into Indian history. The Pax Mogulica has been used
to describe a "golden" age before the fall of the great Empire and
political chaos of the eighteenth century, by Indian historians to
boost national confidence and by British colonialist historians to
justify the Raj as a benevolent attempt to re-establishorder among
the Orientals.
Broadly speaking, however, Mughal historiography has paid
more attention to the brittle foundations and fragile nature of the
so-called peace established by tributary empires. Not only large-
Seevan Nijf 1997: 92-93 on IK I, 74. Robert 1969: 545-548 published this and a more
fragmentary but similar inscription which suggests that the practice was repeated over
several years.
87 The European chartered companies "just as some prominent indigenous merchants"
managed to secure discounts through patronage. Such privilege came at a high price
and doubts were often voiced as to whether they might not have donejust as well by
simplypayingtheunmodifieclrate.ThediaryoftheSurmanembassyto theMughalcourt
in 1714-1716, The Oriental and India Office Records (London), Home Miscellaneous
Series, vol. 69, provides very instructive evidence on the high costs and the numerous
complications and intrigues arising out of attempts to obtain fiscal exemptions. The
fiscal privileges were also often ignored on the ground. Entangled in the shifting ties
of patronage networks, they were frequently being undermined and needed permanent
defence. See in general Hasan 2004, chapter 7 and Prakash 1985: 41-45.
229
 IMPERIAL BAZAAR
but also small-scale local rebellions and other forms of low-level
violence hold a more prominent position in the history of the
north Indian Empire than they do in most periods of Roman his-
tory. It wouldbefoolhardyto claimcategoricallythattherewere
no differences between the two empires in the respective levels
of violence. So many factors of culture, social organisation and
geography affect the scope for successful banditry that one could
never expect exact similarity. Even so the difference can to a large
extent be explained with reference to separate historiographical
traditions rather than genuine contrasts between past worlds. In
the last couple of decades Roman historians have become increas-
ingly aware of the ideological nature of the claim of the Roman
emperors and the indications of persistent violent disorder in the
empire. Though successful for centuries, small-scale rebellions,
occasionally growing to significant proportions, violent disputes
or even minor wars between local communities and raids by moun-
tain people, nomads and "barbarians" were part of the empire in
the bestoftimes. The Roman world wasonly relatively pacified. 9
Banditry and piracy continued to constitute a real menace for
merchants and other travellers, even in Italy.90 This was a fact of
life acknowledged by all, including the imperial state. Roman law
prohibited the build-up of stores of arms intended for rebellion
or social disturbance, but its subjects were still allowed to retain
weapons for their own defence. Merchants and travellers were to
a large extent expected to be able to take care of themselves and
defendtheirpossessionsfrom violent attackers.9 The relationship
between the level of taxation and the quality of protection was
not necessarily very close. An extreme illustration of this is the
caravan trade bringing luxuries from India through the Syrian
desert to Palmyra and from there further on into the Empire. The
88 The fragile foundationof the Mughal peacewas alreadya main theme in Habib 1963.
See Hintze 1997 for some historiographical overviews.
89 Woolf 1993 and Brunt 1990, chapter 11.
90 Cf. Pliny Ep. 6. 25 reporting the disappearanceof a Roman knight duringtravel in Italy.
See Shaw 1984; Hopwood 1989; Griinewald 1999; andWolff2003 (though empiricist,
rather than analytical) for the continuation of'banditry. For piracy, Braund 1993 and De
Souza1999, chapter6. De SouzaandWolffthough, seemtoo impressedwith the claims
of the imperial ideology of Fax.
91 Dig. 48. 6. 1 with Brunt 1990:257-259.
230
A THIN LINE: PORTORIUM, PROTECTION AND PREDAT10N
imperial state claimed the 25% customs duty on this traffic. Still,
theRoman army doesnotappeartohavebeenconspicuously active
in ensuringthe safety of the caravans.Thatwasa task mainly left
to semi-pnvate initiative and the city council ofPalmyra. The rich
inscriptionalevidencefrom the Syrianoasiscity informs us about
the need for local grandees, commanding influence among the
deserttribes, to stepin andensurethesafepassageofthecaravans.
In additionto this, the city seemsto havemaintaineda local militia
which patrolled and stationed the desert routes, and fought offthe
occasionalbedouinattackon city territory. 92
Much has been made of the proximity ofParthia and the politi-
calinabilityoftheRomanarmyto actin the frontierzonewithout
increasing the risk ofprovoking war, to explain the role ofPalmyra
in patrolling the desert. There may be a grain of truth in this, espe-
daily when the laxity on the Syro-Parthian route is contrasted
with the heavier state control exercised in the Egyptian desert on
the roads carrying Eastern luxuries. But the importance of such
considerations can easily be exaggerated. Even in the Egyptian
desert, the state presence seems to have been as much a question
of keeping an eye on the valuable goods to reduce smuggling as
providing protection. The merchants were still expected to hire
their own guards and pay a road tax on them as well. 93 If any-
thing, it is rather the relatively heavy presence on the Egyptian
desert roads that constitutes the exception. The Roman army nor-
mally left it to local initiative to maintain law and order around
the empire. 94 In that respect there is nothing surprising about
Palmyra's semi-autonomous policing of the desert. An inscrip-
tion from the early third century shows the Rhodians organising a
92 Somrner2005:202-213(on the complexinterrelationshipbetweenPalmyraandthe
nomadic tribes of its hinterland, based on cooperation, intermittently disrupted by
armedconfrontation). SeefurtherTeixidor 1984; R. Drexhage 1988;Will 1992; and
Young 2001, chapter4.
93 OG/S674,line12(thoughemended)fordutypayableonaccompanyingguards.P.Vin-
dob. G 40822, recto column 2 fortheuseofprivate guards incommercial transports in
Egypt.Young2001:69-74analysesthedesertroads.Faroqhi1984:52-55offerscom-
parative confirmation from Ottoman Anatolia that imperial surveillance might happen
as much to avoid smuggling as to provide protection.
94 Liebenam 1900: 357-359 long ago noted the sparing involvement of the imperial
government in the policing of its territory. See Wolff 2003, chapter 8 for a recent
discussion.
231
 IMPERIAL BAZAAR
quasi-military expedition to teach bandits and pirates a lesson. 95
The emperor Commodus praised the city of Bubon as a model
worthy ofemulation by othercommunities because"you hastened
with suchgreatenthusiasmto the arrestofthe bandits,overcoming
them, killing some, and capturing others". 96 Much of this locally
organisedprotection will have been rather amateurish, sometimes
ad hoc in character and often of doubtful value. This, for instance,
was the experience of a pair of pig merchants returning from
the village of Philadelphia to Arsinoe in second-century Egypt.
Though the locally employed guard came to their rescue, this did
not deter the bandits from attacking. On the contrary, the guard
simply receivedthe sametreatment asthe unfortunatetraders,who
were tied, given a good beating andhad a pig stolen.97
Protection offered by the different segments of the imperial sys-
tem often left much to be desired. A relatively high degree of
low-level violence was tolerated. Not infrequently one gets the
impression that gangs of bandits were harboured by members of
the local elite, thevery group supposed to underwrite local order. 98
The exaction of customs cannot, in other words, simply be seen
as paymentfor protection. Anyhow, this is not somethingwhichis
peculiarto tributary empires. A pacifiedcivil society is only a rel-
atively recent phenomenon. The important but elusive question is
how the Roman imperial system compares to other pre-industrial
societies. In the sharpest attempt so far to conceptualise the process
of customs exaction in agrarianate societies, the American histo-
rian F. C. Lane suggested that customs dues be seen as comprising
two elements: the costs ofproducingprotection, anda furthertrib-
ute, a monopoly profit exacted by threat of using military force.
95 S£ 1946-1947: 337-338, no. i5&-ni with De Souza [999:218-219.
9<' A£ 1979 no. 624 (translation quoted from Lendon 1997:137) with Wolff 2003: Il8-l2l,
196-197.
97 P. Fay. 108. Alston 1995: 86-96 observes that the Roman state was involved in these
local arrangements. We possess papyri showingdistrict centurions ordering local com-
munities to man watchtowers with guards. But the same papyri attest the crucial role
left to the initiative of local communities in executing imperial orders and the huge
scope for local "negligence". None of this is to deny the occasional imperial assistance
to local policing units, but Alston 1995: 81-86 is too optimistic in his assessment of
the scope of this, as is Wolff 2003, chapter 9. The available manpower resources were
simply inadequate, as is clear from the convenient and still valuable surveys of Sherk
1955 and 1957.
98 Shaw 1984;Hopwood 1989.
232
A THIN LINE: PORTORJUM, PROTECTION AND PREDATION
The stronger the position of the customs exactor, the greater the
proportion of tribute. In other words, fewerprotection-producing
rivals would lessen the opportunities open to merchants for taking
their goods by a different route. The result would be to weaken
their negotiatingpowervis-a-visthe tribute takers andthe cost of
protection, taxation that is, would increase."
The large tributary empires had been created by conquering
tributary rivals and absorbing them into the same system. As we
have seen repeatedly above, it would be wrong to describe these
empiresasunitaryentitiesexercisinga strong,centralisedmonop-
olistic powerin society. The emperors' reachfell far short ofthat.
But the alliance between the different segments of the imperial
state did in some ways reduce, not, I emphasise, eradicate, com-
petition and strengthen the position of tribute takers in the body
politic. The result, asargued in Chapter 2, wasan intensification of
surplus extraction or, to use the language of Lane, rising monop-
olistic profits in the sale of protection. Essentially these could be
realised in two ways, raising the price and cutting costs. It is pos-
sible to argue that both methods occurred in the Roman Empire,
not least in the sphere of customs collection.
The need to accommodate the revenue demands both of the
central imperial state and of local aristocratically ruled commu-
nities would probably have worked towards increasing the total
burden ofcustoms andother dues on trade. 100This didnot happen
overnight The process would have been slow, piecemeal andvery
gradual. Nonetheless, over time from the consolidation of impe-
rialrule underAugustus until thefourth century, sucha trend does
seem to make itself discernible. The central imperial government
graduallyattemptedto confiscatemore andmore local tolls, aswe
have already seen. It also occasionally introduced new dues, for
instance onparticular roadsand,in the fourth century, the so-called
chrysargyron tax to bepaid by trades people, among others. 101 The
local aristocratic segments ofthe imperial state would have sought
w Lane 1966:383-428;Steensgaard1981.
Incidentally Purcell 2005: 222 is open to the possibility of rising customs under the
Romans.
101 We know^ for instance, that Tiberius confiscated many local vectigalia (Suet. Tib.
49.2). Galbaremittedsome.ThiswasannulledbyVespasianwhoneededmoney.He
evenaddednewandheavierones(Suet. Vesp. 16. 1). SeefurtherHerodianII.'iv, 7,
233

IMPERIAL BAZAAR
compensation for some of the lost vectigalia by introducing new
excise duties. Such tolls remained important for financing local
government. 102Taken together, the combined pressure for revenue
exerted by the (competing) demands of central government and
local elite groups, one may hypothesise, would have created a
trend towards slowly increasing the tax burden on trade. 103There
were countercurrents too. Occasionally the emperor remitted or
cancelled vectigalia. He also tried to impose some kind of con-
trol on the number of locally exacted dues. However, as already
argued above, these actions were of relatively limited extent and
usually only marginally successful; the emperor could always do
with more money andhealsohadto placate local allies. The spo-
radic attempts to reduce the burden can therefore best be seen as
precautionary and limited measures intended to counter the risk
ofover-burdening and thus ofkilling offthe trade flows. They did
notreverse the general trend, andwere not even intended to; rather
it was a question of safeguarding the fiscal interest. 104
At the same time, the imperial system probably benefited from
reduced military costs, though only for a period. Conquering and
incorporating many of its neighbours reduced the incidence of
full-scale war drastically in core areas ofthe Empire, such as Italy,
southern France, Asia province, and the Mediterranean Sea. This
meant thatthe effort ofthe army could beconcentrated in thefron-
tier regions, whereas internal policing, aswehave seen, wasleft
very much to the cheaper, irregular "troops" usedby local commu-
nities. They were normally sufficient tokeep violence ata tolerable
who mentions new tolls as one way forward for a princeps in need of extra money.
One senses a pattern of ever-wideningimperial customs exaction This culminated
with Constantine, who laid claim to most local tolls. At around the same time, the
clm'sargynm or collatio lustralis was introduced too, cf. Jones 1964: 431-432 and
871-872. ^ . ..
102 SeeSchwarz200[b: 338-406for a cleardemonstrationof the continuedimportance
of customs in city finances.
.
°3 Liebenam 1900: 359-360 points inthisdirection, too. Buthehadtoomuchconfidence
intheabilityofthecentralgovernmenttoregulatemunicipaltaxation.
I04 As de Laet'1949:463-467"remarkedwe know much less about the formal customs
rates in late antiquitythanduringthe Principate.A mysterious octavahovers n the
background,though"itwasprobablya salestaxratherthana customsduty.Bethatas
it may.ingeneralthelateRomanstateseemstohaveincreasedtaxation.Thereislittle
reasonto thinkthatthis didnot also affecttrade.The introductionofthe chrysargyron
confirms this expectation.
234
A THIN LINE: PORTORIUM, PROTECTION AND PREDATION
level, that is to prevent it from erupting into full-scale war or open
rebellion where tribute extraction would become endangered. As a
result, Roman conquestprobably effected a reduction in the over-
all military burden in the Mediterranean region. Long ago Gibbon
remarked upon the relatively modest size of the imperial army.
Thus the Roman state was able to cash in a peace dividend from
loweredcosts ofproducingprotection in the Mediterraneanworld.
This lasted roughly for the period normally associated with the
heyday of the Pax Romana. By the reign of Marcus Aurelius the
barbarians at the frontiers were beginning to catch up with Roman
military superiority. 105 This made the defence of the Empire more
difficult and military expenditure was again under pressure to rise.
Gradually the peace dividend was eaten up in the coming centuries
by growing costs of the military.
This scenario for the Pax Romana collides head on with tradi-
tional interpretations which have normally focused entirely on the
many benefits of peace accruing to the free movement of goods
around the Empire. But the aim is not to deny that trade would
have benefited from a reduction in the incidence of disruptive wars
in the Mediterranean. The merit of the model, however, is to draw
our attention to the process of how the benefits of empire were
distributed and thus question the automatic, economic textbook
assumption of a straightforward link between peace and grow-
ing interregional trade. Against the benefits of imperial law and
coinage,must also be set the disruptive influences ofthe activities
of the Roman state and its personnel. As one business representa-
tive observed in a letter to "Athenocloros, the merchant", he could
not send the requested grain unless a new boat was sent down
to him because his boats had been requisitioned by the army and
werenot availablefor commercialtransport at thatmoment/06
The significant point to appear from the model outlined above
is that merchants were far from the first in line to benefit from the
imperial order. Here the imperial states really seem to partcompany
with early modern capitalism. Competition for "shares of trade"
between the emerging nation states of Europe forced them into a
different process (see Chapter l). They had to continue to invest
105 Whittaker i994a. '06 BGU XVI, 2644 (26 June, AD 4).
235
 IMPERIAL BAZAAR
more of their income in steadily improving the basic conditions of
tradersbyofferingthembetterprotection,firstthroughmercantilist
privileges, later through firm support for the "free" movement of
goods in the face of local resistance, and finally by effecting a
gradual pacification of their civil societies from the seventeenth
until the nineteenth centuries. 107 The gains from the Fax Romana,
by contrast, were not invested in steadily improving the conditions
of trade still further after its initial establishment. On the contrary,
the tributary state was content to accept the continued existence
of considerable low-level violence. 108 Indeed, its personnel often
contributed to everyday brutality and abuse, as we have seen in this
chapter.Instead,thecentralgovernmentandlandedelitespocketed
theprofitsfromthereducedcostsoforganisingprotectionandwere
probablyeven ableslowly to increasethe price the merchantshad
to pay for this service. In spite of appearances, the conditions of
tradeduringthehighEmpirewerenotnearlyasfavourableasthey
are often imaginedto have been. In the crucial areaof protection,
transaction costs, if anything, probably experienced a slow rise
from the consolidation of the Empire under Augustus until the
fourth century AD. '09
The question still remains, however, whether this slow rise in
protectioncostseveroutweighedthebenefitsderivingfromtheini-
tial reduction in large-scale violence in the Mediterraneanbrought
about by the consolidation of Roman rule. There is no simple
answer to that; we cannot make useful quantitative estimates of
the different forces and balance them off against each other. Only
broad impressions are possible. The key issue is just how much
of a burden on regular trade were war and the concomitant plun-
dering, often referred to as piracy by the ancient authors, before
the coming of the so-called Roman peace. The size of this burden
107 In general Lane 1966, chapters 22-25; Hont 2005 for European state-building and
commercialcompetition;Persson 1999links thecreationofnationalmarketsunderthe
banner of laissez-faire with absolutist governments.
108 Strabo's stray remark (i i . 2. 12) about the negligence ofRoman governors in comparison
to local chieftains in combating brigandage in the more remote corners of the Black
Sea seems no less relevant for conditions in general under Rome than the ideological
celebration of the fruits of peace.
109 Bang 2007 attempts to develop this argument to model the macro conditions of the
Roman political economy from the Republic until late antiquity.
236
A THIN LINE: PORTORIUM, PROTECTION AND PREDATION
in the Hellenistic period is easily exaggerated. Contemporary dis-
cussions of war and large-scale piracy were embroiled in a heavily
ideologised discourse about the true representatives and defend-
ers of civilisation. Military enemies were habitually branded as
pirates and destroyers of civilised order. The Roman conquerors
adopted this symbolic language to set themselves apart from the
alleged unruly chaos of the preceding period of the Hellenistic
monarchies. One should therefore be wary of making the con-
trasttoostark.ThecenturiesprecedingAugustussawconsiderable
commercial developments. Activity in the western Mediterranean
expanded significantly, if we can trust the statistical increase in
the number of shipwrecks which have been discovered, dating to
this period. "0 In the eastern Mediterranean Alexandria, Rhodes,
Delos and the opening up of trade with Indiaall suggest the vital-
ity of commercial activity. Takentogetherthis calls for caution in
estimating the initial improvement in the conditions of trade cre-
ated by the Fax Romana. A revolution in transaction costs would
clearly be much too strong a description of what happened. At
most one can hypothesise moderate improvements which would
then have been subject to gradual erosion by rising tax payments
over the following centuries. In general, the tributary empire only
broughtabout a moderate reduction in protection costs. It did not
fundamentally alter the basic conditions of the bazaar.
If the Roman "peace" nevertheless did lead to an expansion
in the intercity movement of goods, the explanation, as argued
in Chapter 2, can be found in the area of tribute extraction. The
reduction intheincidenceoflarge-scale violence insidetheempire
meant that the process of tribute extraction became less subject to
disruptions. At least for a period, tribute extraction experienced
greater stability and, asa consequence, achieved greater intensity.
The result was an increased demand for the services of the Roman
bazaar. The Empire gave it a greater and, for a long period, more
Parker 1992.The wreckevidencedoesunmistakablyshowa rise in activity during
the last two centuries BC. The changes in the following centuries, however, need not
necessarilybetakenasevidenceof stagnationor decline. Both EgyptianandAfrican
shipssailingtheMediterraneanareunderrepresentedin thematerial,thoughtradeand
transportfrom theseregionspickeduppreciselywhenthematerialbeginsto indicate
contraction, cf. Whittaker 1989.
237
 IMPERIAL BAZAAR
stable flow of resources to handle. '" Such an interpretation finds
support in developments in north India during the dissolution of
the Mughal Empire in the eighteenth century. In areas where the
surplus extracting apparatus remained intact, or even intensified,
trade continued in spite of the increased level of violence. On
the other hand, in regions where the aristocratic, fiscal system
disintegrated, demand collapsed and trade dwindled away. "2
This chapterhasexamined oneaspectofuncertainty intheworld
ofthebazaar:protectioncosts. Theproblem ofsecurity waslarger
than that ensuing from the high-profile phenomenaof piracy and
banditry; it embraced also the no less important element offiscally
motivated customs exactions. However, the merchants and traders
of the bazaar were not as defenceless in the face of corrupt and
abusive customs offrcials as they are often portrayed. Rather, a bal-
ance of power existed between merchant and customs collector;
traders were able to establish a modus vivendi, however fragile,
with the authorities. Their ability to circumvent the toughest col-
lection points limited theroom for manoeuvre left to theindividual
customs official. Too much brutality would underminehis income
in the long term. "3 Mechanisms were in place then that enabled
trade to take place in what to modern sentiments was a"rough'
environment. The next chapter will explore the character of such
mechanisms further.
'" Cf. Wickham2005. "2 Bayly 1983, chapters 1-4.
113 Cf. Dirlmeier 1987on thefrequentcollection points oftransitduesalongthe Rhinein
medieval times. If themanycustomscollecteddidnot kill offtrade,it wasbecausethe
collectors were interested in its continuance. But this does not mean, pace the optimistic,
but unfounded assertion of Purcell 2005, that customs did not burden or limit trade.
They did add significantly to the costs and hence limited potential consumption.
238
COMMUNITY: CULT, COURTS, CREDIT AND
COLLABORATIONIN THE BAZAAR
Poormen ofhumble birth sailtheseasandcome toplaces whichtheyhavenever
seenbeforeandwheretheyarestrangerstothepeopletheymeetandwherethey
donot alwayshavepeople to vouchforthem. Yettheycountonbeingsafeonly
by putting trust in theirRomancitizenship.
Cicero, SecondSpeechAgainstGainsVerres,V, [67
BanarasidasandhisfriendNarrotamdas wenttoBanarasforbusiness. Firstthey
offered prayers to Parsvanathand took a vow to observe the usual religious
practices.Theykeptall thefasts,gaveupgreenvegetables,andthecompanyof
prostitutes. InthebrighthalfofBaisakh(April), Samvat1671 (AD 1614),they
completed the vow andofferedworship.
Banarasidas,Ardha-Kathanak(Sharma1970),p. 107
The first of the epigraphs above is normally cited as proof of the
low social standing of Roman merchants. I am interested in a
slightly different aspect ofCicero's rhetorical figure: the question
of security and uncertainty in the world of the trader, habitually
travelling between different communities. In the preceding chap-
ter the customs collector was used as the ideal-typical example
ofhowthe tributary state didnot manage unambiguously to make
conditions of trade secure. A dirigiste state might be expected to
have exercised strong controls on the movement ofgoods through
all branches ofeconomic life. But that wasnot primarily the cause
of commercial uncertainty in the agrarian empires. Firm com-
mand of the market sector or the bazaarin its entirety was com-
pletely beyond the means of the imperial state. In fact, it was
unable even to curb its own officials. ' The inability of the state
to control commercial life, therefore, is not to be mistaken for
the existence of a de facto regime of laissez-faire. The tributary
Thepriceedictot'Diocletianisa tellingifunintentionaldemonstrationofhowunrealistic
suchcontrol wasunderRomanconditions;its grandiloquenttoneandsheerabundance
ofdetail are signsofweakness,not strength.
239
 IMPERIAL BAZAAR
state simply was not strong enough to provide the level of guar-
anteed security and predictable administration implied by Adam
Smith whenhecoinedhisinfluential adviceforoptimum economic
performance.
Telling confirmation of this can be found in Pliny's Bithynian
correspondencewithTrajan.Thecity ofJuliopolisin theprovince
suffered from difficulties with looting and assault on civilians by
passingarmy contingents. Hence the city had applied to the gov-
ernor to have military personnel stationed to protect it from the
abuses of troops on the march. Apparently Pliny was inclined to
grant the city this favour which was already enjoyed by Byzan-
tium. But Trajan said no on the grounds that it would set a danger-
ous example to the rest of the cities in the province. The guards
in Byzantium were to remain the absolute exception, a carefully
guarded privilege, not a right. If another city was to receive impe-
rial guards, others would soon follow with the same demands.
The result would be an intolerable increase in the administrative
burden of the province. Instead, Trajan preferred to direct his gov-
ernor to do his best to see to it that breaches of military discipline
were punished. This was effectively a decision to leave things
be. Pliny must already have been aware that he could not contain
the military abuses with the normal means available to him. In
practice, benign foresight and care, those much praised qualities
in the emperor, turned out to make little difference to the sub-
jects on the ground. Even Trajan, generally celebrated as optimus
princeps by an admiring posterity, had to allow a certain level of
abuse to continue unchecked.2 Roman peace did not come with
the night watchman demanded by classical economics. The trib-
utary empire lacked both the will and the capacity to fulfil the
role.
The mainproblem ofthebazaartrader wasnottoomuch govern-
mental control, but too little. 3 Mixing in a foreign community or
2 Plin. Ep. X, 77-78. Pace Wolff 2003: 203-204 who cites these letters as reflecting a
processofimproving imperial security arrangementsfor thepopulation. But that means
ignoring the negative answer to Pliny's request. In this case, that is ill advised. The letters
werepublishedandcirculatedas representinga standardworthy ofemulation.
3 See Jones 1988, chapter 8 for the relatively low degree of social organisationof the
market sectorcreatedby the empires. SeefurtherGeertz 1963,chapter3, modifiedby
Geertz 1979.
240
COMMUNITY
engaging^withpersonsofsuperiorstatusorofficialrankposedall
sorts ofphysical as well as financial risks. Merchants and traders.
however, should not be seen as helpless in the face of these chal-
lenges. In the last chapter, their ability to vote with their feet. so to
speak,waj.seentohaveworkedasa counterbalance against'over~-
taxation. But merchant response was not restricted to individual
acts of smuggling andavoidance. Instead the people of the trad-
i-ng^vorld.setaboutcreatinstheirownsocialinstitutions,back-up
mechanisms asitwere,tocompensate fortherelatively weakgua7-
antees provided by the state system. If at all possible most traders
preferred to have "people to vouch for them". 4 Certainly, those
who could do so would not as a general rule have been content to
venture forth into thegreatunknown only trusting in theprotec-
tive powers oftheir Roman citizenship, as Cicero boasted. Instead.
merchantsallovertheRomanworld.justasinotherpre-industrial
societies, attempted to forge personal links and construct social
networks of various kinds which, to some extent, would cush-
ion them against the vicissitudes ofcommercial life, theprocess
Geertz referred to as clientelisation. Two key components of the
social fabric of the bazaar were communal associations and the
household. This chapter examines the role of the former in creat-
ing ties ofcommunity andtrust among traders andofthe latter in
providing basic capital and credit to finance commerce.
Communalassociation- thepowerofreligiousconviviality
ThecopiousepigraphicrecordsurvivingfromtheEmpirecontains
numerous inscriptions revealing the existence of a rich texture of
communal merchantorganisations andotherprofessional bodiesin
many,cities"5 These/ecorcls' incised inmarble andputondisplay,
mainly inform us about the public life of the professional asso
ciations, or guilds. They bear witness to the religious and social
4 Cic. II Verr. V, 167.
Waltzing 1895-1900 remains unsurpassed and collects most of the evidence for nrn-
Ss!onalassodatonsin!heRomanEmpirelnVOIS-3-4,including'thoseofmerchmt^
othercentra:IWOIte arcMickwiu'936; DeRobertis i-97"i':Aus nel"i'982:"De"sZ^
1992;vanNijf 1997.Mommsen 1843-isa seminalcontnbution."""""" "'"" "" "'"
241

IMPERIAL BAZAAR
rituals performed by thetraderconfraternities, collegia in Latin.6
Students familiar with medieval guilds have been struck by the
general absence in their Greco-Roman predecessors of formally
recognised powers to regulate and define the rules of the trade,
suchasnumberofpractitioners,pricesetc. Comparedto medieval
guilds, the professional associations of the ancient world appear
nottohavebeengreatlyconcernedwitheconomicobjectives.Their
focuswasoncommunitymatters, suchasensuringa properburial
of deceased "brethren" or providing a setting for convivial gath-
erings or cultic activities. The distinction between socio-religious
clubs and economic guilds, however, can easily be overdrawn.
Absence of a securely established monopoly on economic regu-
lation, and predominance of "cultural" functions, should not, as
is the case for instance in Finley's work, be equated with eco-
nomic impotence per se, let alone lack of an economic function
for these merchant societies in general.7 Once more, comparison
with Mughal Indiamay prove illuminating.
In many respects, discussions of the role of religious caste in
Indiaresemble thearguments ofclassicists overtheexactnature of
professional associations in the ancient economy. Caste has been
portrayed as a religiously ordered division of labour in society.
The clearest expression of this view was offered by the French
ethnographer Louis Dumont in his Homo Hierarchicus (1966).
In his version caste is seen as a rigid, unchanging system which,
since time immemorial, has organised Indian society into a large
number of hereditary, functionally specialised groups, so-called
castes. Each caste would ideally comprise the members of a spe-
cific occupation; the salt-merchants' caste, the barbers' caste etc.
But caste has in fact never been able to hold Indian society in such
6 Patterson 1994 and van Nijf 1997 treat the public, ritual life of Greco-Roman civic
associations.
7 Seevan Nijf 1997,chapter i fora tentative critique ofFinley's analysis. Rauh 1993 shows
the economic function ofthe religious aspects oftrading guilds at Delos. Finley s analysis
(19853: 81 and 137-138) of the non-econoinic character of Greco-Roman collegia was
basedonWaltzing 1895-1900 via Mickwitz 1936. Waltzing, however, wasmore nuanced
than Finley. He also argued that collegia, though different from medieval guilds, did
fulfil important economic functions: "L'influence et 1'utilite des colleges aupoint devue
economique furent-elles done nulles? Certes non... les negociants et les ouvners...
s'associaient afin de mieux defendre leur interets communs" (1895-1900, vol. I: i88-
189).
242
COMMUNITY
an iron grip. Paradoxically, to the extent that the very hardened,
static and stylised version of the caste system has ever had any
social reality, this seems to have been the result of the changes
Indian society went through after the fall of the Mughal Empire
and in particular during British colonial rule. During this period,
organisation along caste lines was intensified. Caste was increas-
ingly employed by different groups as a means to defend, carve
out and assert their position in a struggle over accessto resources
and privilege in a rapidly changing social order. As a result divi-
sions between different segments of society became more clearly
defined and caste boundaries hardened. Before this period, during
Mughal rule, caste seems to have been much less dominant, far
more flexible and not nearly as highly articulated as was later to
be the case.
The family history of the seventeenth-century Baniatrader and
Jain guru, Banarasidas, provides illustration. Originating from a
clan of so-called Hindu Rajputs, probably peasants, the family
converted to Jainism (a sect bordering on Hinduism) and joined
the caste or sub-caste Srimal. But throughout the history, we find
Banarasidas and members of his family still embracing more
traditional Hindu beliefs and engaging in business partnerships
with members of other creeds and sub-castes. Banarasidas' father
Karagsen, though himself a Jain, collaborated in the jewel trade
with a Siva worshipper and member of the Oswal sub-caste. Unit-
ing these different socio-religiousgroups was a "vague"notion of
the Baniacommunity. The Banias were a very broad and loosely
defined group of Jain and Hindu commercial people (traders,
bankers etc. ). 9 Add to this the co-existence of other similarly
broad commercial groupings, such as Parsis, the Khattris, par-
ticularly in the central part of the Empire, and powerful Mus-
lim groups of merchants, for instance at the hub of the commer-
cial network radiating from Surat using Banias as assistants, and
the image of a clearly defined caste system begins to fall apart.
The social reality appears much more fluid and far less tidy than
the traditional view allows. In economic terms this means it would
Bayly 1999 is fundamental.
y Banarasidas, Ardha-Kathanak, in Sharma 1970: 52 and 56 and Habib 1990:382-383.
243
 IMPERIAL BAZAAR
be wrong to seethe differentmerchantcastesas identicalin func-
tion with medieval guilds. They were more social andreligious in
nature. They did not as a general rule constitute clearly formalised
bodies regulating their trades. Only in cases where a particular
caste came to monopolise a trade locally did it establish control of
commercial practices in ways more closely resembling medieval
guilds. More often castes formed rather loose conglomerates of
numerous locally co-existing sub-castes, usually of an ad hoc
nature, which could also combine to form cross-caste councils of
merchants.
None of this deprived castes of economic significance. Though
not strongly corporate in character, local caste bodies did usually
still influence conditions in the bazaar in many ways and provided
some of the underpinnings and social architecture of commercial
life.Thecommonreligiousidentityproducedbylocalcastegather-
ingscouldoftenbemobilisedbymerchants whenneededtodefend
their interests. In 1669 the qazi (local law-and-orderofficer) ini-
tiatedthe destruction of Hindu temples in Surat, the greattrading
port, following an order from the Mughal emperor Aurangzeb.
The Baniacommunity answeredby closing their shops and leav-
ing the city en masse. The consequence, asreported by the English
factor, was that "the people of Surat suffered great want ... there
was not any provisions to be got, the tanksall and customs house
shut; no money to be procured; soe much as for home expenses;
much less for trade, which was wholy at a stand. "" Only when
the qau agreed to disregard the imperial order and abandon the
policy of religious persecution did the Banias return and city life
was restored to its former level of activity. Religious power was no
trifling matter in the tributary empires. It was a powerful rallying
point. The shared experience of cultic rituals, meals and festivals
created a sense ofcommunity among theparticipants, whichcould
be mobilised in social conflicts.
Greco-Roman collegia were not hereditary castes of course.
But they did seek to create a communal identity around common
participation in regular religious and convivial activities. Time
10 Ashin Das Gupta 1979: 14-15 and 74-88.
'' The EnglishFactories,vol. 13: 190-192,citedin Singh 1972:224. SeefurtherHasail
2004: 4"-43 and 60-65.
244
COMMUNITY
and time again the sources allow us a glimpse of their potential
for social mobilisation. In his speech against the governor Flac-
cus, the Jewish-Alexandrine intellectual Philo describes how the
notorious Isidorus roused the membership of several Alexandrine
confratemities against Flaccus. 12 The potential power of private
associations centred on religious ritual and conviviality should not
be underestimated.
This lesson was not lost on the imperial authorities. Their atti-
tude towards private associations always remained ambiguous.
Though far from identical in their policies, both Mughal and
Romanemperors looked with suspicionon the capacityof various
religiously organised, autonomous groups to turn rebellious, riot
anddisrupttheestablishedsocialorder. 13OneoftheGreatMughal
Aurangzeb's imperial farmans (edicts), for instance, declares that
"men and women visit shrines of saints on their anniversary days,
Friday nights and nights of the last days of the months. This fact is a
cause of disturbance. It should be arranged that none should crowd
at the tombs. '"4 A few lines above in the samefarman the emperor
makes a similar-sounding complaint about a Hindu festival. In
the Roman law codes under a title concerning private associa-
tions we findthejurist Marcianusopiningthatreligiousgatherings
shouldbe tolerated as long as they do notevolve into illegal activi-
ties. 15 This concern is mirrored in Tertullian's late second-century
defence of Christianity directed to the Roman authorities. It is
telling that he chose to argue that the Church should be considered
legitimate by the Empire because it did not present a threat to the
political and social order, in contrast to many other associations.
Furthermore, in strict law professional organisations may quite
12 Philo In Flac. 136-138 and Seland 1996.
13 See Veyne 1979 for the "disciplinary"attitude of the Roman authorities designed to
uphold the moral and social order of society.
14 Mir'at-i Ahmadi, 233.
5 Dig.47. 22. 1. 1 .Thesuspicionthatreligiouslyorganisedbodiesmightturnrebelliousand
go againstthe state is a constant of Greco-Romanhistory, from the Bacchanalianaffair
in early second-century BC Roman Italy right down to the imperialjuristic writings of
the Antonine and Severanperiod. Goodexamplesof the suspiciousattitude are Philo In
Flaccum4 and Trajan's comment to Pliny (Ep. X, 34): "whatevername we give them
and for whateverreason, if people assemble for a common purpose they soon turn into
brotherhoods". See Cotter 1996.
'6 Tertullian Apol. 38. 1-3.
245
 IMPERIAL BAZAAR
possibly have been required to obtain an imperial authorisation. '7
But that was imperial hyperbole. In practice, many associations of
merchants andothercommercial peoplecontinuedto existwithout
imperial blueprint and without interference from the authorities as
long as they did not actually disturb the social order. The govern-
ment had to tolerate their existence, outside the law as it were,
since it had neither the intention nor the ability to take on the
functions performed by these bodies.
In fact, the imperial state andlocal authoritieson occasionactu-
ally needed the cooperation of some of the many associations.
These were of course granted the right to meet. But more than
that, through involvement with state institutions some of these
professional bodies did develop a more consolidated corporate
character of sorts with a formally recognised position in society.
In early imperial Egypt, local groups of merchants bought the
right to administerthe state-imposedmonopoly in the salt trade in
their areas. 19 The municipal fire brigades formed on the basis of
builders' associations in Italy and some cities in other parts of the
Empire are a prominent example.2" A further instance is the pay-
ment of sales taxes and fulfilment of other officially imposed obli-
gationsvia the agencyof some of the professionalassociations.21
However, the development of more consolidatedcoqrorate bodies
was, as far as the sources allow us to judge, slow, uneven andpre-
dominantly, thoughnot exclusively, a feature ofthe late Empire. It
was the product of the expansion of the administrative apparatus
17 Dig. 47. 22. 3. 1 supported by Dig. 47. 22. 1. pr. This explains the habit ofsome professional
collegia of announcing in inscriptions that they had permission to meet, e. g. CIL VI,
4416. Waltzing 1895-1900,vol. I: 114-132 is still the best treatment.
18 A clear indication of this is the admission by Marcianus (Dig. 47. 22. l. pr. ) that people
of small means should be allowed to form burial societies without formal authorisation
as long as they do not meet more than once a month. In practice, however, many
more colleges were tolerated. Waltzing 1895-1900, vol. l: 132-140 clearly brings out
the discrepancy between formal law and administrative implementation. De Robertis
l97l, vol. i: 323-395 and de Ligt 2000, argue against Waltzing, but unconvincingly.
They suggest that the right of association granted to the tenuiores should be seen as a
general permission to form collegia for any puqwse they wished, including professional.
Waltzing 1896, vol. 1: 142-153 correctly rejected this theory.
19 P. Midi. V, 245 with van Nijf 1997: 13-14 and Alston 1998: 175.
20 In Sentinum they were referred to as the tria collegia principalia, CIL XI, 5749. Cf.
Royden 1988 and Patterson 1994.
21 P. Land. Inv. 1562 versa (second century) reveals some groups, e. g. the vegetable
sellers, being taxed with a lump sum payable by the "guild" (line 13).
246
COMMUNITY
and the concomitant deepening of its reach. Until quite recently,
there was a tendency to exaggerate the general extent of the pro-
cess, with historians trusting imperial claims ofomnipotence more
than they do at present. But the most pronounced crystallisation
of corporate features happened primarily in connection with col-
leges involved with imperial supply systems, such as the annonae
of Rome, and the armies in the provinces. None of these associa-
tions were ever allowed to take on a strong independent identity.
Their functions were too important to the state. It kept them sub-
dued as its servants and transformed them in the process. The
shippers, corpora naviculariorum, serving the annona of Rome
and Constantinople were, for instance, changed into an associa-
tion oflandowners, possessing specifically designated lands which
carried the obligation to supply ships for the annona. Hence the
famous hesitations of St Augustine as to whether to accept the
inheritanceof one of those properties on behalfof the Church.22
Thus, even if the tributary empire was never favourable, and
often even inimical, to the formation of strong corporations, it
both hadto accept the existence ofnumerous communal groupings
within the bazaar and at times actually promoted some corporate
features in order to serve its own ends. In the commercial centre of
late Mughal Surat, the need of the state for the active cooperation
of merchants in the administration of local society does seem to
have invested some merchant bodies with more corporate charac-
teristics. 3 Roman professional associations, therefore, were not
necessarily without a measure of bargaining power. Occasionally
the sources allow us a glimpse of commercial groups applying
theircollectivemuscleto someperceivedwrong,muchin thestyle
of castegroups andmerchantcouncils in the MughalEmpire. An
inscription records the decision by a high-ranking official, proba-
bly the praefectus annonae, seeking to accommodate the com-
plaint of the five corporations of shippers from Arles serving the
lt Augustine Serm. 355, 5 with Jones 1964: 827-828; De Salvo 1992 treats the corpora
navicutariorum. Collegia reporting commodity prices, as we encountered in chapter 3,
would also have sharpened corporate profiles. In general, see Mickwitz 1936,
chapter 5. See further Waltzing 1895-1900, vol. 2; De Robertis 1971, vol. 2, parts
4-5, and Ruggini 1973.
3 Hasan2004: 43-49 (should bebalancedby the accountof trouble anddivision in Das
Gupta 1979).
247
 IMPERIAL BAZAAR
annona, over some unspecifiedabuse. A local imperial procurator
is ordered to address and correct the injury in order to ensure the
continued operation of state transports. 4 One should not, how-
ever, see this exclusively as a phenomenon restricted to corpora
involved in state service. To be sure, it is the area best attested in
the sources. But that seems largely due to the character of what
has survived from antiquity. The "private" sector has not left us
many detailedrecords ofcollective commercial life. Most inscrip-
tions, our main source, are very short, concise and predominantly
informative about"public ritual". On the otherhand, the activities
of the semi-official corpora are primarily known from various law
texts, imperial rulings and administrative decisions. It is only to
be expectedthat suchtexts would tend to reflect the main areas of
direct state involvement in economic life.
Even so, a few examples do survive in the records of collec-
tive action taken by professional groups not strongly integrated
intotheadministrativeapparatus.A damagedinscription,probably
late secondcentury, found nearMagnesiabut most probably from
Ephesos, informs us about a conflict with the bakers in the city.
Apparently a deadlockhadbeenreached:"Thepeople areplunged
into disorder and tumults. " As a last resort, a high-ranking impe-
rial official was made to intervene by the city council and bring
"the seditious groups of bakers in the market" to order. 5 A set of
council minutes from Egypt treats of merchants complaining of
and in effect negotiating about the price they are offered for spin-
ning yarn to be used for a consignment of cloth requisitioned by
the imperial government. 2 Otherexamples include linenworkers
in Tarsus, weavers in Oxyrhynchus and builders in late antique
Sardis.27 Though not amounting to the solid privileges enjoyed
generally by many guilds in medieval Europe, these examples do
24 !LS 6987 (second or early third century) with Hobenreich 1997:97-1l5.
25 CIG 2374, translation cited from Broughton in Frank 1933-40, vol. 4: 847. Finley
(l985a: 226 n. 57) wasunduly clismissiveofthe text. But he wasclearly right in arguing
against the vision ot'Rostovtzeff(l957: 178-179) of a capitalist system with an urban
labour class on strike against its capitalistexploiters.
2 P. Oxy. 1414 (late third or early fourth century).
27 Dio Chrysostomus Or. 34. 21-23 (weavers of Tarsus); van Minnen 1986 (Egyptian
weavers); Gamsey l998b, chapter 5 (builders of Sarclis). In general van Nijf 1997:
82-107.
248
COMMUNITY
bespeak the existence of collective organisational power and the
use of communal association by professional groups to defend
their economic interests. This might sometimes even have taken
the form of tacit collective agreements on prices.28 The ambi-
tion to defend and promote their collective interests, however, is
also reflected in the widespread habit of such societies to elect
patrons from among their most prominent members and not least
from variousstrataofthepoliticalelite (local andimperial accord-
ing to circumstances). 29 The relationship of patronage cannot be
reducedto a questionmerely offindingrichcontributorsto cult and
convivial activities. Patrons could be asked to intervene in corn-
mercial disputes, offerprotection in social conflicts andsoften the
attitudeoftheauthorities.30Likewisecollegesmight,forexample,
be employed in various ways to support their patrons in political
contests and conflicts.3'
A web of ties - buildingcommunity
It is possible, however, to focus too much on the exterior role
of collective organisations in the bazaar. Their function was not
simply that of defending the members against the outside world.
Theirnature wasdual. They were at least as important in serving
ascentres ofcommunity, assistingtheformationandreproduction
of social ties. The shared identity constructed around common
participationinculticcommensalitycreateda senseofcommunity
among the membership. It provided a social environment where
28 Thusa small setofimperialrulings,collectedin theCodexJustinianusundernumber
4.59, forbid the collusion ofcommercial bodieson matters ofprice. I seelittle reason
to restrict thepossibilityofsuchinformal agreementsamonggroups ofmerchants and
other professionals to late antiquity.
29 CILXIII, 1688 (an equestrian patron for a guild of shippers); 1900, 1911 and 1916
(municipal councillors as patrons ofcollegia); 1954 (a wine merchant, sevir andpatron
oftheequitesRomaniaswell ashiscollegium); ILS7490(anequestrianoil traderfrom
Baetica and wine merchant from Lugdunum and patron of the latter coqioration). In
general, see Clemente 1972 and De Salvo 1992: 265-287.
30 Van Nijf 1997: 82-iir. The material from Ostia, treated best by Meiggs 1973,
chapter 14, is particularly good. C1LXIV, 4144 (second century); CIL VI, 1759 (AD
389) and 1872 (AD 206) all show patrons actively defending the interests of their
collegia.
31 Asin Pompeii,Mouritsen 1988:65-68and 175;Jongman1988,chapter7.
249
 IMPERIAL BAZAAR COMMUNITY

business could take place. Here traders would forge contacts and
find "people to vouch for them". Nothing illustrates this better
than the habit of merchants, trading in foreign cities, of forming
and using communal associations, frequently structured around
their native identities, in their host societies. These not only pro-
vided shelter in a foreign environment, but also furnished traders
with socialconnections.32 Banarasidas,the MughalJainmerchant
mentioned above, settled in several different cities to do business
during his career. But it is characteristic that most of the time . -v'.--.. ';'. -'.-
he was moving within a network of relatively closely connected
sub-castes of the wider Bania community. It was, for instance, in I i .
'.,
this social setting that he found his frequently changingbusiness ' '. ''-.
associates.33
The networks of traders between as well as within cities cannot '. '-. ^'

be mapped with nearly the same precision in the Roman as in

the Mughal Empire. The epigraphic record, our main source, is
simply too fragmented and contains too many gaps. Alexandrian (a)
merchants might certainly be expected to figure prominently in the
corpus ofextant inscriptions. Several literary sources confirm their
commercial significance. Yet they hardly appear at all. 34 Nonethe-
less, there is sufficient material to allow us occasionally to follow
the community-building strategies pursued by Roman merchants
in some detail. Rome and its harbour city Ostia, the greatest mar- LU 0

ket in the Empire, has produced some illuminating evidence, most 0N

<

importantly in the famous Piazzaledelle Corporazioniin Ostiaof LU a:

^ 0
< Q-
the early to high Empire. There, a number of different "ethnic" N a:

groups of merchants and shippers each possessed a small section Is

a.
of a large portico surrounding a temple. The surviving pavement
mosaics give the names of some of these associations: The Ship- SS5!8StS8te8??':

persandMerchantsofKarales;StationoftheSabratensians"andso
on. Thus we see merchants from distantplaces "building"a social (b)
Fig. 5. 1 Piazzale delle Corporazioni
32 Per. Mar. Eryth. 16, first century AD, provides a clear demonstration of the principle. (a), top, view of Piazzaledelle Corporazioni from the theatre in Ostia
Merchants from Muza in Arabia maintained trade links with East Africa by employing
other Arab captainswho hadacquireda base in the foreigncommunity. (b), bottom, plan of the complex (from CIL XIV, suppl. Ostiense: 662)
33 See Habib 1990: 382-383 on Banarasidas. Markovits 2000 studies one such caste
network from the eighteenthto the twentiethcenturies.
34 It is tempting to see many of them lurking among the worshippers of the Egyptian
cults (e. g. Isis) which spread across the Empire. See Noy 2000: 245-251, though too
cautious, for a discussionregardingthe city of Rome.

250 251
 IMPERIAL BAZAAR

towards the Persian Gulf and India. In addition, Palmyrene com-

munities did at some point exist in, amongotherplaces, Seleucia,
Babylon and Dura Europos, Coptos and Denderah in Egypt aswell
as in Rome (Fig. 5. 2). 41 The vitality of this network was supported
by a rich communal life centred on the religious and civic insti-
tutions of the "mother" city. Temples of their native gods were
established by the Palmyrenes when temporarily or permanently
resident away from home in the foreign cities. 42 Even as far away ^
as Coptos an inscriptionrecords the existence of what appears to 8
be a religious community building belonging to the Palmyrene M

merchants plying the waters of the Red Sea.43 The impressive

catchofinscriptionsrecoveredatPalmyraoffersilluminatingevi- I
dence of the workings of a moral economy inside this widespread u

merchant network. Several inscriptions found in the desert city's 2I

&.§
grandiose temple of Bel honour euergetists for their benefactions
towards or on behalf of Palmyrenes living in the diaspora. One g)'I
Aqqia was commemorated for erecting a temple in Vologesias. s§
m an
Anotherinscription showsthePalmyreneandGreekmerchantsin q

Seleucia setting up a statue to their leader, who had made bene- 11

d Vs
factions to the temple of Bel. Civic and religious ritual, securely
anchored in the institutions of the "home" city, was used as a tls
&
mediumformaintainingandexpressingsocialrelationshipsinside M£
the merchantcommunity even acrossdifferentlocalities.44 Ut '0
w tj

The practices of Indian merchants of the seventeenth and eigh- fc g,

teenth centuries may help further to elucidate the significance I
of such activities. Cultic and social ritual did not just offer a Q

focal point for establishing contacts and bringing people together.

0
&
41 Young 2001: 140 andR. Drexhage 1988,chapter4. SeeNoy 2000: 234-245 for the
Palmyrene presence in Rome.
42 Dirven 1999 has studied this phenomenon for Dura. 43 Young 2001: 8o-8l.
44 Inv. Palm. IX, 6a (Palmyrene and Greek merchants in Seleucia); CIS 11, 3, 3917 (the
tribe of Gadiboli honour Aqqia for a temple erected in Vologesias). See also CISII,
3, 39'6 and Inv. Palm. IX, ll. The practice should not be dismissed as a Palmyrene
exception.Anotherexampleofsuchcommunalties insidediasporasisprovidedbyan
inscriptionfrom Puteoli (OG/S595, AD 174). It records the attempt by the Tyrians
in Puteoli to exploit their connections with Tyre on the Phoeniciancoast to receive
financial support from its city council. Normally the Puteolan station ofthe Tyrians had
been subsidised by the branch of the diaspora residing in Rome. But this had come
to an end. The Tyrian city council, however, refused the request and decided that an
appealshouldbedirectedtotheRomancommunitytoresumesupportoftheirPuteolan
brothers.

254
 IMPERIAL BAZAAR COMMUNITY

!i

ilife
^i^---.. - ; -':. ^'.. "-^.A^®
^^&--L
.

J'^, :< .. :< ?^ '. >. <

v!ff^ .

(d)
Figure 5. 2 (cont.)
'^..: (d), close-up of one of the statue bases with bilingual inscription

-,---, Its effect was much more profound and penetrating. It allowed
the commercial household a means of asserting or improving its
position within the business community. For the ambitious mer-
chant this "meant", to cite C. A. Bayly, "above all, playing an
active and steady part in the temple as well as the bazaar. "45 Life
I '.'. around "the temple" provided a context for merchants to express,

,
al (re)shape and cement social hierarchies within the world of the
bazaar. It was important in stabilising relations and adding sym-
bolic form to an otherwise fluent and opportunistic trading environ-
ment. The universe of the temple offered a vision ofhierarchy in the
world of the bazaar. It was, so to speak, a re-enactment and state-
ment in socio-religious terms of the shape of life in the business
community.
Fig. 5. 2 (cont.)
Such a model seems to capture well the epigraphic habit of
(b), top, view ofthecentral cella (to the right) andthecourtyardofthe
enormous temple of Bel in Palmyra the Palmyrene traders. In the inscriptions of the Bel-temple, the
(c), bottom, section of the Bel temple's portico colonnaded streets and the agora, a very clear sense of hierarchy

45 Bayly 1983: 373.

256 257
 IMPERIAL BAZAAR
among the traders emerges. 46 At the bottom, groups ofmerchants,
most commonly the members of the caravans, pose as grateful
honorants, thanking benefactors for services rendered. Towards
theotherendofthe spectruma selectgroupofcommercialleaders
stand out as patrons of the smaller merchants. They receive pub-
lie honour for assisting the caravans "in every way possible". 47
Thereby their commercial primacy received a symbolic recogni-
tion, onoccasion even supported by declarations ofthe local senate.
They werenotmerely wealthymen in themarket. They werecast
in the mould of community leaders, pillars of merchant society.
This is a role they seem readily to have accepted and reinforced.
The grandees of Palmyra's long-distance trade even undertook to
organise military expeditions to combat intrusive bedouins, who
harassed the caravans. Thus they conformed to patterns commonly
linked with Greco-Roman euergetism. Office holding in the city,
for instance, does not seem to have been beyond the reach of the
top-ranking merchants. Not only did they command local militias,
they also held other civic magistracies in the city. In additionwe
46 How to interpret the hierarchy in the inscriptional record has been a concern of several
scholars during the last century. Rostovtzeff 19326 saw it as evidence of a group of
"merchant princes" in Palmyra. For Will 1957 the merchant princes were not really
merchants. They were more like desert sheikhs using their finances and their control of
vast pastures to supply the merchants travelling in caravans with funding and transport
animals. Though free from the modernising connotations of Rostovtzeff's analysis, the
distinction seems very artificial. The merchant princes of the Italian Renaissance, it
shouldbe remembered, alsocombinedcapitalistactivitieswith aspectsoftheworldof
the knights, as the viewer of Donatello's famous, first statue of St George in Florence
will know. Young 2001: 149-157is a ratherconfusingandat times muddledcritique.
But he rightly insists that the material does not warrant a clear separation of traders
and patron sheikhs. His own interpretation, however, almost eliminating the group
of "patrons" and making the caravan leader the most central commercial figure, is
unconvincing. It is also improbable on comparative grounds. It was normal in "rich'
tradesthattheleading,financingmerchantswouldbesedentarywhiletheirapprentices,
smaller merchants andpeople on the rise wouldbe travelling (van Leur 1955: 62-71;
Goitein 1967: 156-160; and Das Gupta 1991: 355-356). This is the arrangement we
encounter in Palmyra. Inv. Palm. X, 107 shows Akgar, son of the prominent merchant
MarcusUlpiusYarhai,conductinga caravanwhilehisfatherprovidesassistancefromhis
base in Palmyra. The role ofcaravanleaderseemsa sort ofjunior leadershipposition:
the culmination of a career for some; for others a stepping stone on the way to the
very top. Themerchantprinces,asre-emphasisedby Yon2002,mustbeplacedinthe
context ofa complex interrelationship between steppe andcity - they had a foot in each
camp.
47 Inv. Palm. X, 44, 107, III.
258
COMMUNITY
havealreadyseenhowtheypaidfor temples orfor theperformance
of sacrifices.4
No other, or perhapsalmostno other, groupofmerchantswithin
the orbit of Greco-Roman civilisation ever managed to achieve
such a prominent position, including complete integration in city
government. That was a function of the uniquely important role of
the rich Oriental luxury trade in Palmyra's economy. However, on a
more modestscalesimilarclaimsto socialleadershipinthetrading
world were made by the wealthier merchants across the Empire.
The inscriptionspresent them actingas benefactorsandpatrons of
professionalassociations,andin particularas seviriAugustales,of
thepolis at large.Weseetheminvestingin munificence,payingfor
temples, festivals and dinners to the greater good of the bazaar.49
In thatway,ashasrecently beenremarked,commercial groups and
especially their leaders were able to claim a publicly recognised
role within the social hierarchy of Roman cities. 50 This motive
was even more prominent in the status-seeking strategies of Indian
merchants in the Mughal and especially early colonial era. But the
comparison also invites us to look for the economic significance
of such behaviour. In the world of Indian merchants the financial
credit and the social prestige enjoyed by a commercial household
were not easily separable. Failure in "the temple" could easily lead
to serious damage in the bazaar. 51
Such an experience would probably not have been foreign to
merchants in the Roman world. The Latin word for credit was
fides. Fides, however, was also a function of the trust and authority
one was able to command in the sphere of personal relations. The
word was one of the key concepts used to describe the relationship
between patron and client. 52 It was a social, as well as economic,
4 A prominent example is the commercial and civic leader Soadu; R. Drexhage 1988,
nos. 15 and 16 with pp. 101-102. Seealso CISII, 3, 3942.
49 Seviri Augustales: CILXIII, 1960, 1966, 1972; C/LXIV, 3l8 and 397; patrons, benefac-
tions and handouts: C1LXIII, 1954; CILXIV, 4142, 4620-4622; payment for temples
and other public structures CIL VI, 814, 1035 (with Andreau 1987:122-123)and XIV,
2793; Rauh 1993: 289-338.
50 Van Nijf 1997, chapters 4-6. 5I Bayly 1983: chapter 10.
52 Skydsgaard 1976 on fides in relation to the conduct of business and credit. A crisis of
fides might therefore be a "financial"crisis, the collapse or narrowing of credit, Caes.
BC 3. 1. 2; Cic. De Imp. Cn. Pomp. 19. See further Bang I997b and Verboven 2002:
170-182, 287-330and 349-35I.
259
 IMPERIAL BAZAAR
measure of the pulling power and attractiveness of the household
in interpersonal engagementsof all sorts. In their epitaphsRoman
merchantscanthereforebe founddescribingthemselves asfidelis-
simi. Some emphasise how they had always been open to help
friends in need and generous in extending hospitality. Complaints
about the failure of friends to honour their moral obligations are
also voiced. It is in the same context that one should understand
claims that merchants had led blameless, spotless (sine maculd),
frugal (frugi) or "pure" (castus) lives or that they hadlived together
with their wives sine ulla animi laesione, "without damage to the
soul". This was a way of indicating that the household had been
a healthy and well-administered "enterprise"; its members would
make for reliable business associates. 53 Honour was integral to the
conduct of trade. 54
Investment in social standingcannot, therefore, be dismissedas
wasteful conspicuous consumption. It was sound business. Main-
taining a prominent position in the ritual and social life of the
business community served to safeguardor improve thefides, the
credit and honour, of the merchant household. From this perspec-
tive, as remarked by C. A. Bayly, "the distinction between bad
moral and bad economic conduct disappears".55 The communal
activities of the merchants take on importance as a means of pro-
viding the bazaarwith all the trappings of a social system proper.
It was not merely a market. It became a social universe foster-
ing a sense of hierarchy and promoting norms of proper conduct
between individual traders.56 The principle can be seen at work
very clearly in the autobiography, referred to earlier, ofthe Mughal
JaintraderBanarasidas.In thequotationat the start ofthis chapter,
he tells how he initiated a new business venture together with a
53 See ILS 7480, 7518, 7519, 7530, 7531, 7542 for a selection of expressions used by
merchants to describe the moral worth of their households. The link between moral
worth, social status and sound management of the household is a central theme in
Greco-Roman moralising, Cicero Off. 2, 52-64 and 86-87; Plutarch Comparison of
Ansteides and Cato, Horace Senn. I, 2; 2, 3, l66-l8l and Ep. I, 6, 15-42.
54 Plant. Mere. 418-424 and Tab. Vindol. II, 343 connect failure to honour an agreement
with loss of fides and shame. In the Vindolandaletter the gloss erubescam(I will blush)
is used.
55 Bayly 1983: 385.
56 See P. Midi. 243. As emphasised by Hasan 2004: 62, collective organisations combined
economic and social roles.
260
COMMUNITY
partner by performing sacrifices and beginning to observe his reli-
gious duties more strictly. Conduct of business and ritual acts are
overlapping. Thus, he succeeds in depicting himself and his part-
ner as sound, responsible people who could be trusted not to waste
others' money on enjoyments (prostitutes and the like) instead of
making profitable investments.
The world of Roman traders also knew such edifying tales. The
cult of Hercules was popular among traders. A short story about the
Roman merchant Marcus Octavius Herrenus and his relationship
to the cult of Hercules has been preserved for us in late antique
commentaries. Macrobius relates how Herrenus, after having had
success in trade, failed to deliver on his promise of paying a tithe
of his profits to the temple of Hercules. During a voyage Herrenus
encounters a violent storm but escapes miraculously. Afterwards
Hercules reveals to him in a dream that it is he who saved Herrenus
and conquered the storm. Hence Herrenus dedicates a temple in
Rome to Hercules with the epithet "victorious".57 Long-term suc-
cess in business required proper conduct in the world of religious
and civic ritual. The popularity ofthe cult ofthe goddess Fides (Pis-
tis among Greeks) within Greco-Roman bazaars is another very
visible demonstration ofthis.58 The collective life oftraders in the
empire performedanimportant functionin serving as a vehiclefor
the dissemination of norms and the reproduction of social order
anddiscipline among the participants. This can also be seen from
the fact that professional associations were normally organisedas
miniature replicas of a city state with popular assembly and magis-
trates. It was common that members of such societies would refer
to each other as "friends" or occasionally even as "brothers".59
57 Macrob. Sat. 3. 6. 11. Fordiscussions, see Rauh 1993: II6-I20and Coarelli 1988: 185-
204. The catch of inscriptions and altars found at Colijnsplaat at the mouth of the
Rhine in the modern-day Netherlands is a good illustration of such activities. These
were dedications by Roman merchants and sailors to the goddess Nehalenia in the
second/third century AD thanking her for safe passage and return; see Stuart and
Bogaers 2001 for a catalogue.
58 Rauh 1993: I [ [-112.
51j Waltzing 1895-1900, vol. 1: 322-333. Kloppenborg and Wilson 1996 passim. The
organisation of professional associations along city state lines should therefore not
be seen entirely in terms of adherence to the city state, as van Nijf 1997: 68-69 and
chapters 4-6 does. It was also a matter of creating a parallel society.
26l
 T
IMPERIAL BAZAAR
Members who failed to live up to the societal ideals of the clubs
risked being fined or in the worst case faced expulsion.
Courts, arbitration and witnesses
The ancientprofessional associations, in otherwords, constituted
small parallel societies. In that respect they shared many of the
functions of Indian castes. Both helped to give shape to commer-
cial society and create social institutions. One important task was
theadministrationofjustice.Castesservedasjudicialbodiesofthe
first instance. They would normally seek to regulate legal disputes
internally.Onlywhencastegroupsfailedtore-establishsocialcon-
sensus and mediate between litigants was the local Mughal court
approached.61 Any judiciary role of Greco-Roman collegia was
never formally recognised to quite the same extent. The corporate
character ofthe city state wasfarmore clearly articulated in Greco-
RomansocietythaninIndia.Everypolishaditsownlocalcourtand
guarded its political rights within the imperial framework. Only
rarely, such as in the case of Jewish diaspora communities, were
resident subgroups officially granted the privilege of using their
own internal courts, though only grudgingly. 62 Even so, the city
state and the imperial government left much room for semi-private
administration and implementation of justice, indeed, depended
upon it. Both state institutions had fairly weak instruments of
enforcement available for the resolution of private disputes. In
most cases it was left to the winning side of an argument to bring
a favourable verdict into effect. Self-help was always a crucial
element in ancient justice. This made involvement of the "closest'
communityinmostlegaldisputesinevitable.Itssupportwasallbut
indispensableto the litigants anda key factorin making solutions
last.63TheexpectationofPlatoinThe Lawsthatanattempt would
60 P. Mich. 243 and Waltzing 1895-1900,vol. I: 378. See Alston 2002: 207-212 for an
excellent analysisofthe intertwined social, cultural, religiousandeconomic functions
performed by urban "guilds" in the cities of Roman Egypt.
61 Hasan 2004: chapter 6; Bayly 1983, chapter 10.
62 JosephusAnt. Jud.XIV, 185-267andXVI, 160-178withCotter 1996:86.
63 Finley 1983: 1-49 and 118-121. Gelzer 1912 is still unsurpassed on this, and is no less
relevant for the Empire than the Republic (Wallace-Hadrill 1989). Seefurther Lintott
262
COMMUNITY
be made to settle commercial disagreements through mediation
by neighbours or arbiters before approaching the official courts,
remained relevant for all of antiquity (and beyond). 4
As socialhierarchiesgrewsteeperandinternal coherencewithin
the numerous poleis weakened during Roman rule, one may
legitimately surmise that the need for private legal arbitration
increased. 5 But very little is known of these practices. Our evi-
dence tends to overemphasise the role of the state because its
activitiesproduceda significantbodyoflegalwritingsandjuridical
documents.ThebishopsofearlyChristianity,it is clear, attempted,
as part of the care of their congregations, to play the role of medi-
ator and arbiter among quarrelling members of their flock. The
very few ordinances of professional associations surviving from
Roman antiquity, however, leave us very much in the dark about
their part in conflict resolution, and any judgement must remain
conjectural. 7 But there are some strong indications that profes-
sionalassociationsplayeda farmoreactiverole in theprivateadju-
dication of conflicts than is usually thought. If "guild"ordinances
were not normally very detailed, it is still clear that the general
assembly did, at least occasionally,hearanddecideon complaints
raisedby the membership. Oneordinance,for instance,prescribed
that members must not raise grievances about one another dur-
ing communal dinners. Festive occasions were to be kept free of
internal skirmishes. Those should be reserved for treatment by
the monthly assemblies. 68 The regulations of an unknown asso-
ciation from the village of Tebtunis in Roman Egypt, dating to
the reign of Tiberius, add more flesh to the bones. There a fine
is imposed on one who takes legal action against any co-member
1968; Gamsey 1970; Gagos and van Minnen 1994; and Nippel 1995. See Galanter
1981 for an analysis pointing out the critical importance of arbitration, even for the
functioning of modem judiciary systems. With fewer tools of state enforcement, the
relevance can only have been greater in antiquity.
64 Plato Leg. 92oc-d.
65 Cf. Ziegler 1971: 161-163;Garnsey 1970is fundamental.
f>6 See Harries 1999, chapter 10 for how this role continued to shape the character of
episcopal justice even after it had been granted some official recognition during the
fourth century.
67 SeeWaltzing 1895-1900,vol. i: 368-378for a basicdiscussionofthe general assem-
blies of these associations where members would consider "toutes les affaires".
68 fLS 7212: "si quis quid queri aut referre volet, in conventu referat".
263
 IMPERIAL BAZAAR
of the confraternity.69 The collegiumtried to ensure that conflicts
would be settled internally within the collective rather than in a
court setting outside its immediate control.
The aim of such measures was to contain conflicts and prevent
them from escalating to a level where they threatened the social
order among the membership; the prohibition of prosecuting co-
members is mentioned together with a ban on defaming, intriguing
against or corrupting the households of other members. To judge
from Indian examples, such internally organised maintenance of
the social consensus normally reinforced the established order of
things. Lower-rankingmembers and people on the fringes of the
community found it difficult to stand up against more prominent
merchants.70 The episode set in the marketplace, whichPetronius
included in his novel, the Satyricon, suggests that this experience
was not foreign to the ancient world. The three travelling fictional
heroes get embroiled in a commercial dispute with a peasant and
his wife involving a fine gown and an old rag. As things get heated
thewiderbazaarcommunity interferes in theconflict. A prominent
member takes command of the situation. He takes custody of the
disputed objects until a judge can rule in the case, knowing well
that the three foreigners, lacking support in local society, would
probably not risk showing up at the trial. In that way, he hopes to
keep the objects for himself. This is slapstick comedy of course.
Butthecomedydoesnot ariseoutofa deliberatemisrepresentation
of the social mechanisms involved. Indeed, Petronius depends on
exploiting these to produce the constant and incredible reversals of
fortune which underpin the burlesque humour of the scene. Just as
the main characters of the novel believe that the old tunic, hiding a
valuabletreasureonly they knowabout,is againwithintheirgrasp,
chaos erupts. The unsuspecting peasant couple suddenly realise
that the gown put on sale by the three (anti)heroes hasbeen stolen
from the wife in the first place. Immediately they start shouting
(") P. Mich. 243, lines 7-8 with the editor's commentary. P. Mich. 245, the ordinances of
a guild of salt merchants, also indicates internal regulation of disputes.
70 Pearson 1976: 147-149 and Hasan 2004:95-99.
264
COMMUNITY
that they have found the thieves and put the three outsiders in a
rather awkward position. 71
Lifting a conflict out of the immediate community and bringing
it to a magisterial court carried risks of its own, not necessarily to
be taken lightly. However unfair and hierarchically skewed medi-
ation within the bazaar might occasionally become, it was still
often preferable to the courts. Mediation allowed greater scope for
negotiation and the adoption of flexible solutions to conflicts. By
comparison a formal trial was a blunter instrument. With few tools
of enforcement available, an official judge might well handle the
conflict roughly.
Why do you not judge for yourselves what is right? Thus, when you go with your
accuser before a magistrate, on the way make an effort to settle the case, or you
may be draggedbefore thejudge, and thejudge hand you over to the officer, and
the officer throw you in prison. I tell you, you will never get out until you have
paid the very last penny.
So the faithful are admonished in the Gospel of Luke.7 Pri-
vate settlement, as it aimed to restore social consensus, though far
from "fair", normally offered a more lenient solution; the prospect
of reaching a successful conclusion improved significantly if an
agreement, at least in some sense, acceptable to all parties could
be found. Court trials, on the contrary, were the result of failure or
breakdown in the conflict-solving mechanisms of the social net-
works in which the combatants were enmeshed. They represented
an escalation of strife. Harsher means were required to restore
peace in the community. 73
The stakes and the risks were higher in a court trial. But on
average the chances of successfully challenging the reigning dis-
tribution of power in a community probably did not increase cor-
respondingly. Local courts, staffed by judges and juries drawn
71 Petr. Sat. 12-15, 7. See also P. Oxy. 3814 on the importance of local power to assert
one's rights within the community.
7 Luke 12:57-59 cited from Gagos and van Minnen 1994: 44; further pp. 30-48. Their
analysis, though about late antiquity, is equally valid for the Principate.
73 As rulings usually still required the active collaboration of the community to remain in
force, thejudge would often tend to reinforcethe alreadydominant party but now more
harshly in order to ensure his continued predominance. See Hinton 1966, chapters 3-4
for a comparative demonstration of this principle.
265

IMPERIAL BAZAAR
from the ruling groups of a city, could more often than not be
supposed to reflect prevailing patterns of influence. The courts of
governors and other imperial administrators were less entangled
in local intrigue, but also more distant and with an interest, gen-
erally, in maintaining amiable relations with local ruling groups
since imperial administration depended on the active collabora-
tion of these. Add to this that the ancient notion ofjustice required
courts in reaching their decisions to take into account the general
status, standingandcharacterof litigants andthe witnessesappear-
ing for or against them. Lawyers were even expected to play on
matters of hierarchy and influence in presenting a case. 74 From an
individualist perspective, this may seem harsh and unjust, but in
a society where the enforcement of rulings usually required the
active collaboration of communities, finding in favour of estab-
lished relations of power nevertheless represents a form ofjustice:
societal justice. 75
The last point deserves emphasis. Even if courts generally
tended to favour and reinforce existing relations of power, they
still did so within the context of a discussion of right and wrong.
There was thus a chance that occasionally a David might defeat a
Goliath. Greco-Roman justice wasnot simply a "donedeal". 76Not
all contestants, moreover, would have been unevenly matched. So
courts were used, and sometimes to a surprising extent.77 Private
administration ofjustice with its inclination towards the status quo
was vulnerable to silent obstruction and the forces of inertia. In
such cases, an indication of the willingness of one of the parties
to up the stakes and take the conflict to another level might be
just the thingrequiredto break a deadlock. Courts, therefore, need
74 See Saller 1982, chapters 2 and 5 for patronage; Crook 1995 for the importance of
rhetoric in the practice of Greco-Roman law; Jongman 1988, chapter 6 for an attempt,
perhaps overly schematic, to argue that witnesses to documents were normally listed
according to rank and status.
75 Geertz 2000, 175-195 for an example anddiscussionof similar phenomena.
7 Harries 1999: 181-184 makes this point even for the more hierarchical late antique
world.
77 Some papyri indicatea remarkablenumberof grievancesbeingpresented to governors.
Often, however, decisions of higher courts consisted in little more than returning the
case to a local forum. This kind of litigiousness, as Harries 1999: 184 observes, needs
to be understood in the context of the complex strategies pursued by local networks of
power in a system ofcommunally administered justice.
266
COMMUNITY
not merely be seen as an alternative to private adjuclication. Fre-
quently, they seem to have been used to reinforce it. Indeed, the
filing of lawsuits often served to scare an opponent into accepting
private arbitration. 78 Among the Murecine wax tablets are some
documents pertaining to a business dispute between the banker
Caius Sulpicius Cinnamus and his sometime companion Caius
Julius Prudens. The case was brought in front of a judge in Febru-
ary AD 55. A document shows the parties agreeing to transfer
the case to private arbitration. Next in the dossier comes a num-
ber of pronouncements of their chosen arbiter Marcus Barbatius
Epaphroditus, setting time, date and location for hearing the case.
Arbitration could be a rather prolonged affair. The dispute had not
yet been finally settled by March 56. But communications clearly
had not broken down completely. The process of mediation prob-
ably enabled slow negotiations to go on in order to explore the
ground for reaching a compromise. 79
These documents remind us that when we think of mediation
administered by the merchants themselves we need not necessarily
and certainly not primarily focus on the more formal aspects of life
inside the business community such as the assemblies of collegia.
The writings of the Roman jurists discuss the arbiter in relation
to the concept of a vir bonus, a "good man". It was standard to
agree on a person of some standing and moral authority within
the community to act as mediator.80Tojudge from the onomastic
features of Barbatius' cognomen, he belonged to the same liber-
tine business circles in which the bank of the Sulpicii operated.
The formal structure of the social and religious associations of the
bazaar would most probably not have played any direct role in the
actual proceedings. Collegia and other communal bodies should
rather be seen as providing an environment that enabled "good
7 Hasan 2004: 106 shows this for Mughal Surat. Harries 1999, chapter 9 makes this point
forcefully for the late Roman world. But her argument applies equally well to the
Principate.
79 TPSulp. 25, 34-39 and 48 with Camodeca's (1999) editorial comments. Another illus-
(ration ofthis strategycanbefound in the dossierofdocumentscompiled over a number
of years by the Jewish woman Babatha in the conflict regarding her son's inheritance.
Petitions for a court hearing are drafted to cajole the unwilling guardians to enter into
negotiations, cf. P. Yacliii 13-15 with Cotton 1993.
Dig. 17. 2.6 and 17. 2. 76.
267
 IMPERIAL BAZAAR
men" to stand out and be identifiable. The same considerations
are equally applicableto witnesses.Witnesseswhoenjoyed a high
reputation in the bazaar would have been an important asset in
judicial disputes. Their word carried greater weight andpeople of
low standing in the community would have found it more diffi-
cult to get such persons to support their case. Symbolic capital
which arose out of life in the communal associations of the bazaar
played an important part in the organisation and maintenance of
the commercial order and hierarchy.
The structure of credit in the community
In their provision of contacts, symbolic capital, trust and credit,
social hierarchy, rules of behaviour, arbitration and witnesses, the
communal associations of the bazaar made a significant contribu-
tion towards the production of "services" necessary to the activities
of merchants. These would also have included prospective part-
ners in business and a network for the provision of extra capital.
Communal organisation enabled the merchant to access a larger
pool of resources, social as well as economic, than was under his
immediate command. But ties of community were far from alone
in creating the social foundations of business; they overlapped and
intersected with other forms of institutions. The most important of
these was the family and (extended) household.The oikos wasthe
basis of merchant operations.
While social and religious clubs provided an outer shell, so to
speak, for business activities, family and household constituted
the core. Banarasidas, our Mughal ideal-typical merchant, turned
repeatedly to family and relatives to raise his trading capital. 3
Banks and money lenders were certainly present in the bazaars
See Bayly 1983: 375 on how merchants with a low standing found it difficult to get
access to arbitration; Hasan 2004, chapter 6 in general for the social bias of courts of
the Mughal qwi. Gamsey 1970 has treated this for the Roman Empire.
Suchactivitiesare,forinstance,envisagedbytheprovisionsoftheSCdeBcicchanalihus
(S. Riccobono et at. (eds. ) 1941, Fontes luris Romani Antejustiniani, vol. I, Florence,
pp. 240-241), cf. the perceptive analysis of lines 10-14by Rauh 1993:254. Seefurther
his chapter 6 and Gabrielsen 2001.
83 Banarasidas,Ardha-Kathanak,in Sharma 1970:67-68 and72-73.
268
COMMUNITY
of India. 84 But their capital was normally of modest proportions,
confined as they were to household organisation,just as with the
merchants. Deposit banking was far from unknown. The small
scale of the overall enterprise, however, imposed strict limits on
the capacity of the banks to turn such deposits into working capi-
tal. They were always vulnerable to insolvency. Most loans from
banks, though far from unimportant, were therefore short-term.
They can perhaps better be described as financial services together
with money assaying and changing, giro transfers and bills of
exchange, which constituted the core operations of Indian banks
and money lenders in relation to the trading world. Such services
were not at all peripheral. They eased the workings of commer-
cial life considerably, for instance by facilitating money transfers
between localities. But they did not constitute the foundation of
trading capital. That came from the more or less extended house-
hold. 8-5
One common procedure, also practisedby Banarasidasand his
father, was to devolve a portion of the household fortune onto
a young member of the family or poorer business relation in a
sort of partnership. With few means of his own, the contribution
of the junior partner consisted of his work. He was expected to
increase the capital outlay by conducting business, frequently on
long trading voyages. The profits were then to be shared in some
way specifiedby custom andpersonal agreement. Suchasymmet-
rical business partnerships were well suited to the uncertainties
and irregularities of the bazaar world. By parcelling out the cap-
ital, the household spread its risks. It also acquired a convenient
and above all fairly reliable means of setting up branches of the
family business in different cities. By employing persons who
would regularly be dependent on the family, and furthermore, by
giving them a stake in the financial outcome, the household tried
to ensure that its representative would remain loyal to its interests.
This avenue also opened an outlet for surplus capital otherwise
4 Indian banking is the most complex example of a pre-modern bankingsystem. It is
therefore a useful benchmark for Roman banking. SeeJain l929;Habib 1972; Goldsmith
1987; Goody 1996: 72-74 and 94; and Haider 1997, chapter 4.
85 Goitein 1967: 161-185 and229-266, though dealing with themedieval Jewishtrading
community in Cairo, provides a useful description ofthe structure ofcredit in the bazaar.
269
 IMPERIAL BAZAAR
stored in the bountiful coffers of the Mughal elite. Members of
the imperial aristocracy sometimes invested or speculated in trade
through merchants connected with their great households. But
it is important not to exaggerate the significance of this. Aris-
tocratic involvement never amounted to commercial dominance.
One can probably better describe it as a process whereby tenta-
cles spreading outwards from the larger households occasionally
penetrated the world of the bazaar without ever taking it over. 6
Instead a picture emerges of a market sphere with many small
individual units, comprising poorer and junior members - often
travelling - of the merchant community. A considerable portion
of these would have been connectedwith larger, more established
households consisting of senior merchants, richer financiers and
the odd aristocrat. 7
A different but equally asymmetrical form of credit extended
by the financing segment of the trading world was the bottomry
loan, used to finance trading voyages. In return the borrower had
to put up the cargo or sometimes the ship as security. If the ship
was lost at sea the money was lost too. Thus the main risk was
borneby the lender whoreceived very high interestpayments on a
successfully completed journey. Rates of 15, 20 or 30% cover the
normal range. Often such loans would be accompanied by quite
detailed instructions for the trip regarding destination, route and
duration.It wasalsocommonforthelenderto senda representative
to accompany the borrower and keep an eye on the investment.
In effect the bottomry loan was another way of expanding the
business of the household without actually having to increase the
size of the firm. Instead, the financierrelied on semi-independent
agent households. Merchants of more equal means also regularly,
when need arose, pooled their resources in partnerships where
risk and profit were shared more equally. Though sometimes of
long duration, most of these were only short-term and frequently
changingengagements. They never really effected a separation of
the trading capital and the individual households. 88
86 Blake 1991: 109-110 and 117-118, modified by Chatterjee 1996: 66.
87 DasGupta 1991 and Raychaudhuri [99[.
88 Banarasidas, Ardha-Kathanak, in Sharma 1970: 71, 107-110, illustrates the transient
nature of partnerships.
270
COMMUNITY
The pattern of credit outlined above is sometimes loosely
referred to, with a phrase borrowed from the Italian Middle Ages,
as a "commenda"-system.s9 Commendawasthe term usedto des-
ignatea particularform ofpartnershipwhereone entrantprovided
the capital and the other the work while profits were shared on
a three-quarter to one-quarter basis. In a common variation of
the arrangement, funds were contributed two-thirds to one-third,
with the junior partner providing the work and receiving half the
profits. 90 Ever since Weber devoted a now classic study to this
phenomenon, commendahas been closely linked to an evolution-
ary story about the rise of capitalism in northern Italy. In Weber's
analysis commendawas treated as a first step towards achieving a
separationofhouseholdfundsandbusinesscapital. The legal form
given to commenda contributed to a development which eventually
enabled jurists to conceptualise the trading capital of merchants
as a fund distinct from and independentof the household, in other
words as a firm in the modem understanding of the word. 91
One needs to make a distinction here. Weber was interested
in how the various specific legal formulas, which developed in
the Italian city states, were gradually manipulated and shaped by
jurists to formthe modemconceptofthetradingfirm. In that sense
commendamay havebecomepart ofa process thateventually saw
the first institutional shoots of capitalism appear in Italy. But in
itself there was nothing intrinsically modern or "developed" about
the commercial practice of commenda, as Weber was the first to
admit. 92 The researches of Udovitch and Goitein in the 1960$
identified varieties of commenda in Jewish, Byzantine and espe-
cially Islamic law that pre-dated the Italian practices.93 Used as
a convenient shorthand for the flexible, and frequently asymmet-
rical forms of finance described above for India, commenda-style
financing seems to be a common and very old expression of the
household-based economy of the bazaar. Our steadily expanding
database on ancient Mesopotamian trade leaves little doubt that
89 VanLeur 1955: 62-73. See further Goitein 1967: 170-183.
c)" Lopez and Raymond 1955: 174-184 for a convenient list of illustrative examples of
medieval Italian cornmenda contracts in translation.
91 Weber 19240. 92 Weber 19240:335-339.
93 Udovitch 1962 andfurtherUdovitch [970 andGoitein 1967.
271
 I
IMPERIAL BAZAAR
the familiar medieval repertoire of partnerships, bottomry loans
and commenda had strong roots in remote Middle Eastern history.
These forms of finance were not in themselves the first step in a
development away from the oikos, but rather reflected its needs
for spreading risks and flexible adaptation to a shitting environ-
ment. 94 The Roman world seems to offer confirmation of this; and
the story ot'Greco-Roman commercial credit needs to be fitted into
this broader picture. The structure of Roman commercial capital
provides a fairly accurate reproduction of the main features of the
pattern outlined above based on the bazaarsof Mughal India.95
Commercial loans were generally short-term, rarely extended
for more than the duration of a single operation. The archive
belonging to a group of Sulpicii freedmen bankers resident in
Puteoli, serving as a harbour for Rome under the Julio-Claudians,
offers a glimpse of the situation in what must have been one of the
busiest markets in the Empire. 96 Two grain traders, Caius Novius
Eunus and Lucius Marius lucundus, are attested raising money
on the security of a couple of medium-sized shipments of mainly
Alexandrianwheat. Eunus secureda loan of some 13,000 sesterces
to be repaid immediately on the request of the lender, but only by
handing over control of goods at least 3-4 times the value. lucun-
dus obtained 20, 000 sesterces for a two-month period by giving
a consignment of goods probably worth twice or three times as
much into the custody of the lender. 97 Short timespan and high
securities are two of the features which are already familiar from
the foregoing discussion of the character of credit in early modem
India. They are a reflection of the close interrelationship between
94 Van Leur 1955: 73; Trolle Larsen 1967 and 1976; Goody 1996: 64-69 (in spite of the
modernist emphasis); Renger 2003.
95 Cf. the reflections offered by Andreau [ 999: 151. Whatfollows runs counter to Temin
and Kessler 2007, but they mistake the societates publicanorum, the companies of
tax-farmers, for merchant companies and ignore the fact that these societates were
exceptional and to a large extent phased out with the advent of monarchy; they cannot
do service as a Roman version of the charteredjoint stock companies of early modem
capitalism.
9 See Camodeca 1999 and 2003 for an edition and analysis of the archive. Though
descriptive and mainly derivative of Camodeca's work, Jones 2006 provides a useful
introduction and survey of the evidence.
97 TPSulp. 45 and 5 1-52 (Eunus), 46; 53 + 79 (lucunclus).
272
COMMUNITY
the need for a flexible and opportunistic approach to business
and the concern to reduce the risks so characteristic of bazaar
households.
Another common feature is the high number of actors and the
relatively circumscribed nature of individual commercial capitals.
A good illustration is provided by a second-century letter sent
from the Alexandrian banker Marcus Claudius Sabinus to two
merchants from Askalon. 98 The document records the payment of
a loan of 7 talents 5, 160 dr. on behalf of two Roman citizens to
the two merchants together with their two further companions and
joint shipowners. The sum total is far from negligible, about half
theminimumcensusfor a city councillorin a wealthycommunity
and enough to procure some 150 tonnes of wheat - an above-
average grain cargo, but somewhat less than the capacity of the
large Alexandrian grain freighters plying the route to Rome. The
merchantspresumablyhadmorevaluablegoodsinmindsincetheir
ship was a relatively small merchantman. Whatever the precise
natureofthetradingventure, it is highlyrevealingthatthefunding
of such a mid-range operation still saw two lenders sharing the risk
and required four merchants to combine their resources to carry
the burden.
Family and extended household partake as important elements
in all ofthe above examples from Roman business life. The Sulpicii
bank of Puteoli was based on a succession of freedmen relation-
ships. Caius Novius Eunus, the freedman trader, maintained rela-
tions to his old master, Caius Novius Cypaerus, who ran a ware-
house where his former slave stored his goods." Ofthe four mer-
chants from Askalon, two were brothers. The painted inscriptions
on Spanishoil amphoraerecoveredfrom MonteTestaccioin Rome
offer some of the best evidence on the household dynamics ofmer-
chantenterprises in the Roman Mediterranean. 100Frequently one
98 SeeCasson 1986foranedition, translation, commentary andanalysis. Seefurther the
discussion by Rathbone 2003: 217-2[9.
l» TPSulp.45.
'°° It matters little whether thestate wasultimately behindthese imports. Private personnel
conducted the actual transports of oil, at least until the third century. The material is
published in CIL XV and updated in Rodriguez-Almeida 1984: 222-233. De Salvo
273
 IMPERIAL BAZAAR
encounters partnerships based on a shared connection to the same
household, such as "Corneliorum" or "Fadiorum". '01
Occasionallyit is possibleto follow the operationof such"busi-
nessfamilies"in greaterdetail. Take for examplethe "DD.Caecil-
iorum". '02 In AD 149 potsherds document D. Caecilius Matemus
operating on his own. D. Caecilius Hospitalis is also on record
trading individually. But in AD 154 the two appear in partner-
ship. At this point it is necessary to introduce a feature particularly
prominent in Roman society - slavery. One might speculate that
both or perhaps one of the two were former slaves. In that case,
we seem to be faced with a situation where the one or both as
slaves were provided with capital to set up individual branches of
the household. As they proved their worth, they were eventually
set free. But it is clear that they still kept close contacts with the
mother household, as evidencedby theirbusinesscollaboration.
However, both the cognomina, Maternus andHospitalis, were only
rarely used of slaves. Maternus in particular seems to have been
common in the Spanish provinces, while the chance of Hospitalis
being used to designate a slave is marginally bigger. 103 Another
possibility, therefore, is that we are looking at the collaboration
of two freebom male members of the same family, or a mixture
of statuses. At any rate, freeborn or libertine, the combination of
members working individually with occasional collaboration in
partnerships between branches of the extended household is char-
acteristic of the flexible patterns produced by a commenda-type.
system. The presence of slavery modifies, but does not fundamen-
tally alter the basicprinciple. Sometimes, ofcourse, suchcollabo-
ration also included business contacts from outside the household.
1992: 183-225 and 255-256 shows the fluidity of partnerships in the Monte Testaccio
material. See further Liou and Tchernia 1994.
.
0' C/L XV, 3844, 3874.
102 CILXV, 3764-3795 for the DD. Caeciliorum. Hospitalis in 3764, Matemus in 3765-
3766 andjointly in 3768-3781.
103 De Salvo 1992: 219-222 and 251 thinks of traders like Hospitalis and Matemus as
belonging to the Spanish municipal aristocracy. Liou and Tchemia 1994 puncture
much of the prosopographicalguessworkwhichhas attempted to identify the traders
appearing on Dressel 20 amphorae with members of the landed elite in Baetica. The
names Matemus and Hospitalis are discussedby Kajanto 1965: 79-80.
274
COMMUNITY
Another inscription sports a business partnership of "DD. Caecil-
iorum et Aelii Optati". '04
The provisionsofRomanlawaddconfirmationofthevalidity of
the impressiongainedfrom the Spanishoil amphorae.They recog-
nise several varieties of the asymmetrical and parcellised form of
capital provision characteristic of commenda and related commer-
cial instruments. °5 First, there is the businesspartnership proper,
the societas. This was conceived as a very flexible instrument by
the Roman jurists. '06 The partnership might be comprehensive or
merely cover a single business venture. It was not generally treated
as giving rise to an independent economic entity, except in excep-
tional circumstances, but was seen mainly as regulating relations
betweenpartners ratherthan with the "outside"world. Therefore it
was consensual. If consent was withdrawn, the partnership ceased
to exist. By the same token, the death of a partner saw the disso-
lution of the partnership. From a modem perspective, this kind of
partnership looks like a fleeting and transient entity. But from a
bazaar perspective, what meets the eye is the adaptability of the
instrument which allowed economic actors to adjust quickly to the
hazardsandopportunities offeredby a shiftingmarket. Oneyearit
might prove expedient to be on your own, while in the next, greater
advantages were to be had from pooling resources with one of your
business connections, as we saw the Caecilii do above. This kind
of partnership also facilitated diversification of risks. One might
explore opportunities in one line of trade or market together with
one partner, while perhaps employing another part of the house-
hold capital in a completely different range of products with other
partners or on your own.
4 CIL XV, 3795; further 3951-3952 and 3973. Another example of business household
and branches is given by Curtin 1991: 90-96 and 167-169 in his discussion of the
Pompeian producer and merchant of fish sauce, Umbricius Scaurus.
105 Verboven 2002, part 3 is an excellent attempt to understand the institutions of Roman
"business law" as reflecting the honour and trust-based personal relationships of the
household, rather than the needs of proto-modem commercial firms. Crook 1967,
chapters 6-7, is still useful on commercial life and the institutions of law.
106 Gaius Inst. 3. 148-154andDig. 17.2. ForsomemoderndiscussionsofRomansocietas
see Johnston 1999: 106-107; Zimmerman 1990, chapter 15. Arangio-Ruiz 1950 is a
classic.
275
 IMPERIAL BAZAAR
The Roman societas, in other words, corresponds quite well
with the flexible patterns we have come to expect from our com-
parisons and shares many of the formal properties characteristic
of partnership in other commenda-type systems. It also spans a
wide spectrum of social relations ranging from complete equality
to stark hierarchy. Among the many possible varieties, it should
by now be unsurprising that the jurists recognised something akin
to a clear example of commenda. It is stated that a partnership can
exist between people where one brings the capital and the other
the work. Additionally, profits may be sharedby agreement while
loss is wholly borne by the capital investor. '07 In other words, the
arrangement described by thejurists is a commenda in all but name.
The only thing missing is that the contract has not been severed
from the general law of partnership and treated under a separate
heading. But that is a matter of legal formalism, not commercial
practice. 108
Roman law elaborated in greater detail a number of other legal
instruments which, like the unequal partnership, enabled the mas-
ter to diversify his interests and devolve part of his capital onto
junior members or associates of the household. Most important
of these were the regulations concerning agency or business man-
agersandthe so-calledpeculium}09Thepeculiumdenoteda share
ofthe householdresources given over as a personal fund to people
in the power of the head of the household, for instance sons and
daughters or slaves. This could be used much like a commenda
107 Dig. 17. 2. 29. pr. -l; 17. 2. 5. 1; 17. 2. 30 and Gaius 3. 149. Santucci 1997 has recently
emphasised the importance of this institution. Reinhardt 1968: 24-68 presents a dis-
cussion of how Roman law came to recognise unequal partnerships. He portrays it in
gradual evolutionary terms from Republic to high Empire. This seems to me mistaken.
The principle is clearly attested already from the late Republic. If jurists continue
to discuss it, this is due precisely to the unequal social relationship underlying the
practice. The more powerful partner would inevitably be tempted in some cases to
challenge the equality of his junior partner. Similar conflicts appear in discussions
of commenda by Muslim jurists, cf. Udovitch 1962: 205 and more extensively 1970,
chapters 3-6.
Ios Cf. Andreau 1999: 151.
109 Gaius Inst. 4. 69-74 remains an exemplary introduction to these instruments. The
relevant text in the Digest are: Book 14 for agency and business managers. Dig. 14.1
(actio exercitoria); 14. 3 (actio institoria); 14. 5; Book 15 deals with the regulations
concerning thepeculium. Some modem surveys and discussions: Kirschenbaum 1987;
Aubert 1994, chapter 2; Wacke 1994; Johnston 1999: 99-108. The relevant chapters
of Buckland 1908 are still useful.
276
COMMUNITY
arrangement. In Plautus' comedy The Merchant, the father invests
his son with a peculium for a trading voyage. On his return the
son is to render accounts, pay back the principal and a share of
the profits. The remainderis for him to keep. 110Many ofthe basic
characteristics of commenda are clearly present here.
Nevertheless, the peculium is more frequently discussed as a
way of achieving the limited liability of modem companies and
separating the business from the household. By making his depen-
dants do business on the basis of a peculium the master would not
be responsible for more than the size of this fund, it is claimed/"
That is not quite what was envisaged by the Roman jurists. They
distinguished clearly between cases where the pater familias had
either ordered his dependants to conduct a particular business
transaction or put them in charge as managers of a shop or busi-
ness enterprise, and cases where their dependants acted on their
own. Only in these latter cases would the liability of the household
be restricted to the size of the peculium, otherwise liability was for
the full amount, in solidum. As Gaius explained in his introduc-
tion to Roman law, if a person had entered into a transaction with
a dependent person authorised by his master, he could certainly
always sue him on thepeculium. "But no one would be that stupid",
he went on, because the master was liable for the full amount irre-
spective of the current holdings of the peculium. "2 A ruling by
the emperor Alexander Severus upheld this view. Even if a person
was liable up to the amount of a peculium, the emperor observed,
he would still be liable for any enrichment he had drawn from the
transaction or for the full amount if he had put his dependant in
charge of the business. 113
Plaut. Mere. 80-97. At least this interpretation seems the most economical. Plautus
describes the arrangement in terms of the father equipping the expedition and counting
out money to the son and giving him instructions about the minimum sale price. The
excess profit is described as a gain to the son's peculium. But the original capital
should undoubtedly be seen as having entered his peculium as a debt to the father. It
wasthroughthepeculium that a son in potestate wasenabledto participatein business,
cf. Johnston 1999: 99-101.
'" E.g. Johnston 1999: 101-105;Zwalve 2002.
"2 Gaius Inst. 4.74, cf. Dig. (Gaius) 14. 5. 1.
"3 Codex Justinianus IV, 25. 2. Cf. Dig. (Ulpian) 14. 1. 1. 22. A slave forming part of the
peculium of a son in power, still renders the pater-familias liable in full if the latter had
authorisedthebusinessmanagementofthe slave. Businesspursuedwith apeculium, in
other words, did not per se limit the liability of the master to the extent of its contents.
277
 IMPERIAL BAZAAR
The logic of the relationship between the head of the household
and his agents appears with particular clarity from a case discussed
by the jurist Paul. A slave had been put in charge of lending out
money against security by his master. On his own volition the same
slave had also begun to offer his services to grain merchants. He
would take over their payment obligations towards their suppli-
ers. Only now, the slave had run away and one of the suppliers
demanded payment from his master as liable for the conduct of
his business manager. The master denied responsibility since he
had not authorised the line of business giving rise to the dispute.
During the trial, however, it emerged that the master had under-
written the activities of his slave in a range of other areas. The
prefect of the Roman grain supply, therefore, decided in favour of
the claimant. On the appeal, however, the legal councillors were
inclined to find in favour of the defendant pater familias since he
had not given a mandate for this particular line of business. But in
the end, the emperor decided to uphold the sentence because the
master had seemed "in everything to have regarded the slave as
acting in his name". "4 There is no question here of regarding the
conduct of business through slave agents in general as based on a
legally sanctioned form of limited liability. 115
The limited liability ofthepeculium must be placed in a different
context. It was an attempt to protect the master from receiving harm
from his slaves. One of the advantages ofcommencla proper is that
it gives the agent a stake in a positive outcome of the business.
As shown by the Plautine example a similar result could also be
achieved directly through the primary activities connected with
the peculium. A complementary method was to allow the slave,
next to his main operation, to do a little business on the side that
would enable him to earn money with which eventually to obtain
his freedom. This seems to be the phenomenon underlying the
"4 Dig. (Paul) 14. 5. 8.
"5 The inadequacy of the focus on limited liability also emerges from Andreau's work.
On several occasions he has discussed the cause celebre involving the future pope
Calistus. According to Hippolytus (Ref. Omn. Haes. 9. 12.l-l 2) the imperial freedman
Carpophorus had set up his slave Calistus in banking. When Calistus defaulted and
fled, his creditors came to the master Carpophorus. In 1999: 67-68 Andreau took this
as a case of the limited liability of the peculium. But now (2006: 205-208) he has
changed his mind and takes it as a case of full liability for the business manager.
278
COMMUNITY
Pauline case. But when the slave was acting on his own initiative,
it was important to protect the master from people trying to exploit
the situation andthereby inflicting a loss onhim. 116That ishowthe
limited liability of the peculium fits into economic practice, not as
a generalised means of investment with limited liability. This is not
to say, of course, that during conflicts masters would not sometimes
manipulate the rules to their advantage and try to claim limited
liability. Inevitably they would, as is clear from the claims which
sparked the two imperial rulings just treated. But such claims had
no secure backing in law, nor in business practice, it would seem.
That is the very clear impression conveyed by the documents in
the archive belonging to the Sulpicii bankers. They provide some
of our richest information on the employment of slave agents
and managers in business. A recurrent feature is the specification
that a particular activity takes place at the master's command or
authorisation. "7 The rules governing the use of business agents
and the peculium reflected the need of households for a flexible
set of instruments through which to delegate the management of
' Dig. (Gaius) 50. 17. 133 states the general principle that masters should benefit from
their slaves, not receive harm. The jurists had little sympathy for those who inflicted
loss on other people by exploiting, colluding with or taking advantage of their slaves.
Instead they emphasise the duty to act responsibly and with circumspection in dealing
with other people's slaves active in business, cf. Dig. 14. 1. 1. pr; 14. 1. 7; 14. 1. 9. Slaves
acting on a peculiuin, therefore, could not give give away presents or stand surety
unless it was clearly in the interests of their master, cf. Dig. (Ulpian) 15. 1. 3. 5-9. For
an analysis, see Kirschenbaum 1987: 27-28 and Buckland 1908: 213-216.
E.g. TPSulp. 45, 46, 48, 51. ?8. Cf. the discussion in Jones 2006: 138-142; the
evidence for authorisation is much clearer and more copious than the vague tit-bits
he can assemble for independent action with liability limited to the peculium. The
last practice can only be hypothesisecl (not more than that) in a few instances, often
of doubtful interpretation. Thus TPSulp. 25 contains the record of a hearing in front
of a Puteolan judge between the Sulpicii and one of their business associates Gaius
Julius Prudens concerning an unpaid debt. In the record Sulpicius asks Prudens to
confirm that he is the owner of two named slaves. Some years before, Prudens had
mandated the Sulpicii to transact with these slaves in his own name (TPSulp. 48). Now
from his reply has been stricken the addition that the said slaves were in his power
"suaque im [sic] potestate". This was to avoid a "noxal action", that is a criminal trial,
Jones explains, with the commentary provided by the editor of the tablets, Camodeca,
and ensure that the case was treated as a civil suit, e.g. an ordinary action on the
peculium. The actio de peculio, however, is not the sole alternative, some of the other
middleman institutions could equally well be implied. But here Jones strangely ignores
Camodeca's commentary: "Sempre nello stesso ordine di idee di progressi rapporti
commerciali fra Ie parti si e pero preferito piuttosto pensare all' actio qiiod iussu o
all' actio institoria" (1999: 86). Avoiding the noxal action might equally well lead to
Pmdens being sued for the full amount, in solidum.
279
 IMPERIAL BAZAAR
its different economic activities/branches to junior members and
dependants (children, slaves and freedmen). In that respect, they
served some of the same functions as a commenda arrangement
proper. 118
Finally, the bottomry loan was a well-established practice in
Roman commerce, too; andthe form it took fits well the argument
advancedhere abouta generalisedasymmetrical, commenda-sty\e
pattern of finance for the bazaar. "9 It was common practice for
the lender to have a representative accompany the borrower on
the trading voyage, thus reducing the latter almost to the level of
a junior partner. One case discussed in the Digest by Scaevola,
an advisorto Marcus Aurelius, saw a slave representative situated
in Berytus advancing a maritime loan to a merchant for a trip to
Brundisium andbackagain.A slave, Eros, was sentto accompany
and supervise the borrower on the journey to make sure he abided
by the terms of the agreement. If for some reason, the merchant was
unable to start his return trip before the end of the sailing season,
the loan was to be repaid immediately. Eros, the accompanying
slave, would then bring the money to Rome where, presumably,
themasterofthetwo slaves wasliving. Behindthisoperationstood
a financier sufficiently rich to maintain a branch of his household
in a foreign harbour. 120
It is worthwhile emphasising the flexibility and adaptability of
these instruments. Bottomry loans wereemployedto financecom-
mercial ventures of very varying scale. The 47, 160 dr. advanced
to the four merchants from Askalon wasgiven as a bottomry loan.
In that case, the lenders do not seem to have had representatives
118 The suggestion by Johnston [999: l02-lo8 that masters essentially left the initiative
to their slaves as to how to invest the household resources, in order to enjoy limited
liability from a position of passivity, is implausible. Roman morality prescribed that
the paterfamilias did not neglect his household and kept a firm hold on the activities
of his subordinates (e. g. Cato Agr. II and IV). A master unable to control his slaves
and freeclmen was an object of condemnation and of ridicule (e. g. Sen. Apocolocyn-
tosis); he was one of the stock characters of Roman comedy. Roman jurispmdents
expected masters to take care when setting up slaves in business. Thus UIpian (Dig.
I4.3-1 [-4) insistedthat responsibilityremainedwith the master if an institor/agenthad
intentionally damaged the owner in conducting business with an unsuspecting third
party.
9 Plutarch Cato Maior 21, 6 and the so-called Muziris Papyrus (Casson 1990; Rathbone
2001).
120 Dig. 45. 1. 122. 1. Two recent discussions, Sirks 2002; Rathbone 2003: 2(6.
280
COMMUNITY
to monitor the movements of the borrowers. Credit on an entirely
different scale is attested by a papyrus detailing the specific con-
ditions of another bottomry loan; it was extended to a merchant,
probably in Alexandria, for a round-trip to Muziris in India to
procure highly prized Oriental luxuries. The value of the cargo
which served as security for the loan amounted to some 7 million
sesterces, 6-7 times the minimum census requirement of a Roman
senator. To safeguard the investment, the lender had agents sta-
tioned at several nodal points along the route in Egypt. The terms
specify that the borrower place the goods under the seal of either
the lenderor his agentsat the warehousein Coptos while awaiting
shipment down the Nile and again on arrival in Alexandria. In the
preceding damagedsection of the papyrus, there may be mention
of further representatives located in the Red Sea harbour where
the cargo would arrive from India and of a camel driver to take the
goods through the desert. With that sort of economic and organ-
isational muscle, it should have been possible for the lender to
conduct the trade on his own account. However, by financingthe
venture of another, he gained one important advantage: he ensured
that the man actually travelling to a foreign and far distant location
had a significant interest in a successful outcome. The travelling
merchant,on theotherhand,gainedthe advantageofhavingsome-
one with stronger shoulders carry the ruinous and very real risks
of losing the capital outlay on the way, to pirates, bad weather or
some other accident. 12'
Thus, allowing for small culture-specific variations, the stmc-
ture of Roman commercial capital can be seen as an example of a
general bazaarpattern which the path-breaking historian of trade
in the Indian Ocean, J. C. van Leur, described for the Indian Ocean
in the sixteenthandseventeenthcenturies andsaw stretchingback
121 Some basic analyses (with varying emphasis) and text: Casson 1990 and Rathbone
2001. Further Rathbone 2003: 220-221; Young 2001: 54-58 and De Romanis 1996,
chapter 4. The suggestion by Rathbone that the papyrus describing the terms of the
loan is a general master contract, a generic document which would then serve as basis
for drawing up loan agreements for specific trading voyages, is no more than a possible
hypothesis. The fact that an account of the specific cargo was written up on the versa
after the completion of the journey, seems to me to speak against the inteqiretation of
the recto as a "generic"rather thana specificdocument.
28l
 IMPERIAL BAZAAR
atleastto Hellenistic times. I22It wasa pattern characterised bythe
predominance of household organisation, fragmentation of capi-
tal in various commenda-like arrangements and transient partner-
ships,butalsobythepresenceofsomeverysubstantial andwealthy
financiers. I23This,ofcourse,alsoopenedanavenueforoccasional
aristocratic involvement at the financing level, as in India. There
hasbeenmuch debateabouttheprecise extent ofsuchactivities. '24
Many contributions havebeenill guided. Considerable aristocratic
involvement incommerce hasbeenclaimedbythecritics ofFinley
as a sign of the economic modernity of the Roman world. 125 But
no plausible level of aristocratic interests in trade could possibly
call into question hismainpoint, thepredominance ofagricultural
Rathbone2003: 214-215 and 225-227, on the contrary, postulates a decisivebreak
withHellenisticpractices,inspiteoftheinstitutionalcontinuities,andpreferstoplace
the Roman phenomena in a context comparable to thejoint stock companies of the
seventeenth andeighteenth centuries. Thatseems excessive andbasedonanargument
from exceptions. The anecdote told by Plutarch of Cato the Elder(Cato Maim 21 .6)
whois allegedto havegroupedfifty merchantborrowerstogetherin a partnership
before lendingthem money, shouldnot be seenas the norm. Quite thereverse. The
number fifty is a round number, chosen foreffect rather thanverisimilitude, andisused
byPlutarch toillustrate Cato'sexcessive andextraordinary zealinmoneymaking. The
otherpieceofkeyevidenceadduced,a rulingpreservedintheDigestona bottomryloan
(22.2.6),ismisunderstoodbyRathbone.Theattemptbythelenderto claimsecurity
in goods financed by other lenders on other ships does not reflect the existence of a
partnership(andnopartnersarementioned).Thepointisthattheborrowingmerchant
had spread his risks by sending his goods on different ships financed by different
lenders. Now, thepresent lender hadclaimed whatremained afterrepayment ofthese
otherloansasanadditionalsecurity. Whenthecargoofferedasprimarysecurityto
him waslost, he tried to reclaimhismoney by arguingthat the loss of the cargo,
in effect, only amounted to a diminutionof its value andthat the borrowertherefore
had to make up the shortfall from the other pledged securities. This, however, was
rejected by thejurisprudents since the cargohadbeenlost underthe terms ofthe sea
loanandtheriskaccordinglyrestedwhollywiththelender.Neitherofthebottomry
loansattestedinthepapyri,incidentally,supportRathbone'sview.Thatinvolvingthe
partnership offourmerchants only represented a very modestconcentration ofcapital,
whilethe enormous, andexceptional, wealthimpliedby the documentconcerning
the commercial trip to Indiaonly involved one lender and one borrower. It is not an
important distinctionthatsomeRomanfinanciersofbottomry loanshadrepresentatives
travelling with the lenders; that practice remained common also in the Middle Ages
andintheIndianOceantradebeforethejointstockcompanies.
' The lastpoint merits emphasis since it is often wrongly claimed thatvanLeurdenied
thestructuralimportanceofbigfinanciermerchants(cf.hisanalysis1955:191-227).
SeeColin 2000for a discussionofsedentaryandtravellingmerchantsin theRoman
world.
SeeGamsey 1976;seeWallace-Hadrill1991 for someofthebestcontributions. See
furtherD'Arms 1981;Picket1983and 1984;andSchleich1983and
D'Arms 1981.
282
COMMUNITY
wealth in the composition of elite incomes. In fact, the dominant
trend within Indian studies has been going in the exact opposite
direction. Here most scholars have been keen to downplay the
extent of involvement by the vast householdsofthe political elite
in commercial activities. They prefer to portray it as a marginal
phenomenon precisely in order to assert the vitality, strength and
potential modernity of the traders of pre-colonial India. '26 That
is a curious contrast in the strategies of argumentation. The same
phenomenon is ascribed the opposite meaning within each field
of scholarship in order to make similar claims about the relative
modernity of commercial life. The contrast cannot be explained
merely as a reflection of a real difference between the worlds of
the Roman and Mughal lords. In neither empire does the evidence
permit a sufficiently solid and detailed reconstruction of the phe-
nomenon to enable such precise comparison. The contrast, in short,
is a product of different historiographicaltraditions.
This is an indication that the terms of the discussion are
wrong. Occasional aristocratic commercial investments. the extent
of which must in any case be supposed to have varied signif-
icantly with the opportunities offered by widely different local
circumstances, do not represent a departure from the general
shape of the bazaar. '27 Indeed, the great concentration of eco-
nomic resources in the hands of the aristocracies in each empire
makes some level of involvement almost inevitable. Wealth on a
scale implied by the cargo from Muziris could be mustered by
only few men within the Roman Empire; and of those, the over-
whelming majority will have belonged to the political elite. In
the archive of the Sulpicii freedmen bankers, one can actually
observe a trickle-down of wealth from the imperial household into
the business community of the harbour city of Puteoli. '28 On the
otherhand,the Indiancomparisonmay serve asa warningagainst
exaggerating the level of aristocratic dominance within the com-
mercial life of the Empire. The Muziris cargo, after all, cannot
I2(' Chatterjee 1996,chapter3 contraBlake 1991: 117-118.
27 SeeBayly 1983:391-393fortheabilityofthebazaarto absorbtheinvolvementofthe
political elite.
128 E.g.TPSulp.49,6^,69;Camodeca2003.SeeRathbone2003:219-223forthealmost
certain elite context of the Muziris cargo.
283
 IMPERIAL BAZAAR
easily be taken as representative of trade in general. The sheer
concentration of monetary value represented by this transaction
necessarily implies that there would have been very few of these
in any given year. Even so, some very wealthy merchants, rather
than landed aristocrats, are likely to have existed in the centres
for the richest and busiest trades. One thinks of the small group
of merchant princes known from Palmyra. Below these heights
there would have been many more operating at a relative distance
from the households of the political elite.I29 The same banker's
archive which shows money finding its way into business from the
emperor's patrimony, also indicates the existence of a commercial
community by no meansgenerally dominatedby aristocraticmas-
ters. 130The most likely scenario, therefore, would see aristocratic
investments going into commercial activities next to those con-
dueled by various more independent groups consisting of richer
merchants andfinanciers- the majority ofwhom wouldhavebeen
of middling wealth. '3' Suchinvolvement, in fact, blends in rather
well with the commercial patterns of uncertainty and opportunistic
distortions by local monopoly groups established in Chapter 3. 132
Merchants and commercial ventures of the great Roman house-
holds would have been able to exploit the political muscle of the
supporting aristocrat to gain special advantages. They might, for
instance, have been more difficult to tax than their competitors
who lacked the support of a grandee; or they might have found it
easiermomentarily to cornera specificline oftradeby mobilising
129 Some inscriptional examples of very rich and merely wealthy merchants: CIL VI,
29722; X, 1872; XIII, 1942, 1954, '96o, 2448; XIV, 4142, 4620-4622 (possibly);
A£ 1973, 646. D'Arms 1981, 140-148 recognised that for a large number among
the wealthiestfreedmen, the so-called seviriAugustales, it was not possibleto claim
aristocraticbacking,Abramenko 1993takesthistoo far; it is inconceivablethathardly
any freedmen Augustales would have owed theirprivileged position to their connec-
tion with an aristocratic household. These households were the major slave owners,
thereforealsothemainmanumitters in Romansocietyandpotentially providedslaves
with better opportunities for enrichment.
130 E. g. TPSulp. 13-14, 53, 78, 8o. See the balanced treatment of Camodeca 2003 (the
Sulpicii,for instance,mayhavehadsomedistantconnectionstooneofthegreataristo-
cratic families, but as we meet them, they seem to have been operating independently).
131 See also Andreau 1999, chapter6 for a discussion of such intermeciiary groups of
independententrepreneurs and financiers.
132 See Prakash 1985: 29-34 and 232-234 and Ali 1997, chapter 6 for Mughal nobles
using their political muscle to .gain from trade.
284
COMMUNITY
the great political and financial power of the noble household to
achieve their end. 133
Thus far the basic structure oftrading capital in the Roman world
appears to correspond with our image of the househo\d-commenda
pattern of the bazaar with many individual actors, fragmentation
of capital and a restricted layer of richer financiers. In terms of
financial services, however, the Roman system does not appear to
have developed quite the same array of instruments as existed in
the bazaars of Mughal India. The availability of forms of insur-
ance was greater in India. But in particular the Indian merchant
had access as a matter of routine to an institutionalised system
of bills of exchange and cheques known as hundis. These were
negotiable, in principle. In practice many obstacles reduced the
extent to which this was the case. The most important function
of the hundi was to enable cash-free paper transfers of money
betweenlocalities. 134Alongthemostfrequentedtradingroutes the
hundi-system operated quite well. Roman bankers and money
lenders never established so routinely institutionalised a system
as this. I35
Nonetheless, the contrast can be exaggerated. Our sources for
Roman banking operations leave many gaps. Much is simply
unknown about the detailed activities of Roman bankers. But
from more recent research it is becoming increasingly clear that
bankers may have been of greater service to traders than is often
thought. 13 For instance, payment not infrequently happened via
a bank account. Thus traders were able, though absent, to receive
and effect payment through their bank representative. Giro trans-
fers are also a possibility one should not dismiss out of hand.
Many banks seem to have been local, but some bankers had cor-
respondents or branches outside their home base which enabled
them to conduct credit and money transactions across geographical
locations without necessarily having physically to move coin. 137
.
33 C. 77i. XIII. I, 5, 7, 15, 2i. Whittaker 1993. chapter 12.
134 Haider 1997: 185-189; Habib 1972.
35 De Ligt 2002, though he much exaggerates the capacity of the Mughal system, cf.
Richards 1981.
13 Contra Biirge 1987andFinley 198511: 141. Andreau 1987is fundamental.
137 The insistence by Rathbone 2003 and Harris 2006 on the importance of various kinds
of non-cash transfers in the Roman economy is surely correct. But their observations
285
 T
IMPERIAL BAZAAR
Finally the evidence from Roman Egypt, recently treated by Hider-
ato, deserves mentioning. A fair number of papyri document the
use of a sort of non-negotiable bill of exchange or cheque, a so-
called epitheke, to create short-term credit and effect paper trans-
fers of money between different locations in the province. 138 It
would be rash to elevate this into a widely used system of bills
of exchange. However, it does remind us that we risk underesti-
mating the intercity transfer of money by credit instruments in the
Roman world. Though clearly unable to match the Indian hundi,
the Roman financial instruments do not belong on an entirely dif-
ferent planet. Perhaps they are better viewed as an earlier example
of a system which gradually reached maturity and became more
institutionalised during the Middle Ages andearly modem period
across Eurasia. '39
Compartmentalised networks of trade
This chapter has focused on the social relationships of Roman
traders, in particular those shaped by the household and commu-
nal associations. The aim has been to demonstrate how the com-
mercial order depended on the ability of its members to form ties
of community. These, both inside or outside the household, were
of crucial importance to the success of the individual merchant.
Whereverhe went, the merchantwould try to use hispersonal net-
work in order to safeguard and promote his interests. A business
letter, dating to second- or third-century Roman Egypt, illustrates
the point:
Ammon to Dionysius his brother, greetings... Provided you know that we could
find a letter of credit at your place in order to pay the money there, I am willing
to come to you to pay the wool. I think the other dyers, too, will go there for
this purpose. When they come, please deal with them on my behalf and write
(and tell me) which one you want me to remit the money to. Many salutations
from your mother Plousias and your sister Hermione and Amoitas your father
need to be balanced by the evidence which suggests a more limited role, cf. de Ligt
2002. For some cautious observations, see Andreau 1987, chapter 18; Andreau 1999:
42-44; Bogaert 2000; Camodeca 2003.
138 Hiderato 1999-2000 developing Preisigke 1910.
139 VanLeur 1955:71-73; Abu-Lughod1989;andGoody 1996,chapters2 and3.
286
COMMUNITY
and Pathermuthis your brother and Pallas. Salutations to the father from myself
and the aforesaid. We pray for the health of yourself and all your household all
your life long. '40
A dyer, Ammon, contacts his business partner living in a differ-
ent location about possibly travelling to the latter's place to buy
wool. However, to do so, he needs to know whether he would
be able to obtain an epitheke, the Egyptian letter of credit, with
which to make payment. Furthermore, Ammon requires his partner
Dionysius to handle his dealings with other dyers coming to the
market and inform him to whom he needs to transfer money. Thus
we see Ammon mobilising his personal connections to forward
his business venture in another place. Without such connections,
he would not be able to conduct business there, or only at much
greaterrisk. Note, therefore, howthe relations withthe partnerare
invested with greater and more enduring meaning by the use of
affectionatemetaphorsto createa senseoffamily, even thoughthe
two were not "brothers", strictly speaking.
Personal contacts, or clientelisation as Geertz called it, were
all-important in the world of the bazaar. They helped to shelter the
merchant in an uncertain and risky market environment. But such
contacts worked through exclusion as well as inclusion. Traders
outside the social network of personal contacts were at a disad-
vantageor even excluded.In economic terms, the variouspersonal
and communal networks served to lower the transaction costs of
insiders while creating higher barriers to entry in the market by
outsiders. Thus communal ties and clientelisation operated to turn
the uncertainties and risks of the bazaar to the advantage of the
traders by reinforcing the fragmentation of business. The result
was a tendency towards the compartmentalisation of commerce.
Trade would often have been conducted within relatively closed
circles and goods would regularly move in separate or parallel
channels. 14
140 P. Oxy. Hels. 48 (trans. the editors). Hiderato 1999-2000: 94-95 treats the letter in
relation to epitheke.
141 Greif 1993. Cf. Geertz 1979.
287
 T
IMPERIAL BAZAAR
This is a phenomenon which can sometimes be documented to
a surprising extent from the archaeological record. *42 Italianterra
sigillata, that is mould-made and stamped red-slip pottery from
the late Republic and early Empire, may serve as illustration. Sur-
viving in substantial quantities, enjoying a wide distribution and
with finds having recently been collected in a database of more
than 35, 000entries, the distribution patterns ofthis type ofpottery
can be examined in unusual detail and enable us to go beyond the
merely impressionistic. 143 The evidence points to the importance
ofdifferentnetworks for the distributionofthe products ofdiffer-
entpotters andproductioncentres. A goodexampleisprovidedby
two of the larger manufacturers of Italian terra sigillata, P. Cor-
nelius andL. Gellius. Bothenjoyeda roughlyequaldistributionin
Italy. Theproducts of Cornelius, however,wereexportedin much
greater quantities to Spain and Morocco, whereas Gellius figures
far more prominently in north Italy and along the Danube. I44 Like-
wise, the products bearing the stamp of MURRIUS, a potter ten-
tatively located in Pisa, are found particularly in NarboneseGaiil,
Hispania Tarraconensis and Africa Proconsularis. By contrast, the
contemporary Pisan products ofCn. Ateius Euhodus also find their
way in substantial numbers to the provinces of GermaniaInferior
and Superior. 14-5 Euhodus was one of a number of Ateii potters,
most of them Pisa-based. Together they make up one of the most
prolific groups of producers. Yet their products are not particu-
larly prominent on Italian sites whereas they seem to have been
exported in large quantities, not least to the Gennanic provinces.
This points to a more general difference between productions in
Pisa and Arezzo further inland. The former seem to have been
part of a commercial network focused on the north-west Meditcr-
ranean littoral, the Gauls and the Gennanies. Arretine products, on
theotherhand,dominatedtheItalianscene,butwerealsoexported,
142 SeeReynolds 1995: 139andLund 1999for some examplesofthecellularmovemeiit
of goods. Bounegru 2006: 86-88 indicates different domains for different Roman
commercial groups in the Danubian and Pontic area.
143 Kenrick 2000. 144 Kenrick 2000:46-49.
145 See Kenrick 2000 for statistics for potter 292 (Ateius Euhodus) and 1202 (Murriii.s).
However, even with [43 and 122 attested finds respectively, the level of statistical
uncertainty in the analysis of individualpotters is still quite high. A single big finil
couldinfluencethe overall shapeofdistributionpatternssignificantly.
288
COMMUNITY
presumablyviadifferentchannels,to a largenumberofprovincial
locations. '46
In the case of the Pisan productions, the legions stationed on
the northern frontier will have been an important market. Simi-
larly closed and separate circuits of supply havebeen suggested in
relation to deliveries made of Spanish olive oil to the Rhine army
and as a model more generally for how military units procured
their supplies. 147 In the analysis presented here, such activities
need not necessarily be taken as anomalous, state-induced devia-
tions from proper markets. As strategies they were not exclusive
to the state. '48 Rather, they blend well into the broad weave of
personalised ties and social networks which have been described
as crucial for the functioning of trade in general. In a bazaar envi-
ronment, lacking the institutional supports developed by capital-
ism, the use of such strategies was an expression of how markets
worked.
146 Thedifference,therefore,cannotsimplybeexplainedintermsofgeographyandthe
easier transport afforded by the location of Pi.sa much closer to the sea. Arretine
products also travelled widely. Geography and social networks interacted to pro-
duce specific patterns. That phenomenon has also been documented for the distri-
button of the various provincial productions of sigillata in the Gauls and Germa-
nies of the second and third centuries AD, see Rapsaet 1987 and Rapsaet-CharIier
1988.
147 Remesal-Rodn'guez 1997 (though the evidential base is weak; the numbers are too
small to make the statistical analysiscertain) andWhittakerl994a: 104-113.
148 Cf. Lo Cascio 2006 insisting that the state was not simply standing outside "the
market . As the biggest landowner, it would have engaged in markets as one, albeit
very powerful, seller ofagriculturalproduceamongothers.
289

EPILEGOMENA
TAKING STOCK - THE WORLD OF GOODS
A French missionary who had viewed these palaces found "it incredible how
rich this sovereign [i. e. the Chinese emperor] is in curiosities and magnificent
objects of all kinds from the Occident". Yet the foreign treasures were only part
of an assemblage that aimed to make the imperial retreats complete with every
imaginable creation of nature as well as humanity. As Granet says, even things
that no collector could find nevertheless figured there, sculpted or drawn: it was
a universal collection of "evocative singularities". Such diversity was directly
linked to the ruler's power.
Marshall Sahlins, Cosmologies of Capitalism, p. 432
Here is brought from every land and sea all the crops of the seasons and the
produce of each land, river, lake, as well as of the arts of the Hellenes and
barbarians, so that if someone should wish to view all these things, he must either
see them by travelling over the whole world or be in this city [Rome]. It cannot
be otherwise than that there always be here an abundance of all that grows and is
manufacturedamongeachpeople. .. Whateverone does not see here neitherdid
nor does exist.
Aelius Aristeides, To Rome, Or. XXVI, 11-13'
The blatant flattery and undisguised hyperbole of the second-
century Greek orator, Aelius Aristeides' eulogy to Rome and
her Empire jar with modem sentiments. Students of the Roman
world are now increasingly convinced that Aristeides could not
in earnest have subscribed to his rosy image of empire. They
have started searching for chinks in the rhetorical armour of the
Greek master: anything to suggest the resistance of a cultured
Hellene to his Roman oppressors. 2 Nonetheless, to generations of
Roman historians the image conveyed by Aristeides' speech has
come to embody their view of the second-century Empire. 3 The
Translationadoptedfrom CharlesBehr (Aristeides,The CompleteWorks, vol. II, 1981)
andJ. H. Oliver (1953).
2 See Swain 1996: chapter 8 for a balanced discussion of Aristeides.
3 Schiavone 2000, chapters I and 2 draws attention to the central place of Aristeides'
speech in Roman historiography.
290
EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
Aristeidean paradigm, so to speak, reached its fullest fruition in
the work of Rostovtzeff, for whom it was the story of the creation
of a liberal, enlightened monarchy that unified the Mediterranean
into one vast economic space - a world market and cosmopoli-
tan bourgeois, capitalist society. Rostovtzeff's construction was
a fragile one. Even in his own analysis the bourgeois order was
constantly threatened by collapse into oppressive and destructive
despotism. Our notion of empire has been caught up in this duality.
A 'primitivist' conception focuses on the exploitative and preda-
tory character of empire. 4 'Modernists', on the other hand, have
conjured up a sort of ideal Smithian market economy.5
Both dimensions in our image of empire are, in fact, present in
Aristeides' speech. If the Roman Empire appears as an agent of
economic integration, it is not exactly in the liberal, laissez-faire
sense, which Rostovtzeff envisaged. It is rather as the omnipotent
commander of the world's resources. The city of Rome is pre-
sented as an irresistible magnet gathering all the globe's variety in
the imperial centre. If it is not to be found there, it does not exist,
as Aristeides hyperbolically claims. This is Rome as ruler of the
oecumene, a universal empire commanding, collecting and con-
suming the world's diversity. Such is the Empire's resourcefulness
and strength that whole regions are laid bare by its demands. In
one ot'Aristeides' oratorical flourishes, the Arabs are depicted as
having to travel to Rome and beg for incense should they wish to
enjoya shareofthatluxuriousproductoftheirowncountry. 6 Inthe
universe of the Greek orator market trade becomes an expression
of imperial submissionandtributary obligations. The characterof
Roman trade is Janus-faced. Both aspects of the imperial experi-
ence, greater flow of resources and exploitative command, must
be accommodated within the same framework in order to achieve
a satisfactory understandingof markets, trade andthe economy of
the Roman Empire.
4 See Finley 198511: 182-183 for instance emphasising pure exploitation in response to
Keith Hopkins' taxes-and-trade model.
5 See Persson 1988, chapter 5 for an interesting example of an economist suggesting
parallels between Rome and the later integration of the European economy.
6 A different variation ofthis ideological theme isdeveloped in Pliny's Panegyric to Trajan
29-32. There it appears in the claim that even Egypt has come to depend on Rome for
its grain. The flow of economic resources is perceived within a context of tribute.
291
 T
EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
It has been the aim of this study to construct such a framework -
not, however, in the form of a simple middling compromise. A
portrait of the Roman world as either a "tailed" version of moder-
nity or an agrarian economy approaching modernisation would
be entirely unsatisfactory. Either of these unconvincing alterna-
tives would leave us with an unattractive conglomerate, neither
this nor that, and would fail to bring us past the current dead-
lock in the debate about the Roman economy in general and trade
in particular. We need to rid our discussions of the implicit and
sometimes explicit modernisation narrative undergirdingmuch of
the current disagreement. To escape our conceptual impasse we
need to turn to comparative history, and what is more, a kind of
comparative history which has never before been attempted. We
need to identify different analytical contexts which resemble the
Roman situation more closely than the early modem European
world that is normally used as point of reference (cf. Chapter i).
Mughal India is one such possibility. Other pre-industrial empires
would lend themselves to the same exercise. But the historiog-
raphy of the Indo-Muslim empire is particularly well suited to
our purposes. The image of the Great Mughals presiding over a
vast and expanding empire based on the floodplains of India has
many features in common with the way Roman hegemony has
been understood. However, whereas the liberal, benevolent con-
ception of Rome hasbeen slow to die, Mughal studieshavehadto
struggle with an equally persistent hardcore vision of an exploita-
tive, Oriental despotism. The notion of a possible modernity has
only rarely possessed the same plausibility to Mughal historians
as it has enjoyed with students of Rome. Their task has there-
fore been slightly different. It has been a question of finding an
appropriate characterisation for a society in which a considerable
transfer of economic resources between regions takes place within
an assuredly, or at least more assuredly, non-modem context. This
has produced an understanding of trade andmarkets that is of use
to us.
In Chapters 2 to 5 I have attempted to use examples from
the Mughal Empire, occasionally with the inclusion of additional
material from otherplaces,to suggestideal-typesor models forthe
shape and character of markets and particularly intercity trade in
292
EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
the Roman Empire. Rather than observing Rome in the traditional
perspective of later European developments, I have been looking
at the Empire through a Mughal-Indianlens. Perhaps in the pro-
cess Rome has come to look more like seventeenth-century India
and seventeenth-century India more like Rome than was actually
the case. This may well provoke some classicists and historians
of Indiaalike. It is a widespreadmisperception that the individual
characteristics of a particular culture are always to be considered
paramount. But cross-cultural similarities should not a priori be
seen as any less real than differences. Differences and similarities
must be handled pragmatically. Whetherto emphasise one or the
other depends on our questions and the problems we examine.
The first chapter argued that, in spite of the obvious cultural con-
trasts, thepolitical economyoftheRomanandMughalas well as a
number of other tributary empires differed in the same fundamen-
tal way from the developing world/state system of early modern
Europe.
This shared difference in the political economy of the two
empires has been made the basis of the analysis. Comparisons
with Mughal India have enabled the identification of a number of
key processes, developments and characteristics which similarly
shaped the economy of the Roman Empire in important ways. This
is not to deny that there were important differences between the
Indo-Muslim and Greco-Roman worlds. There obviously were.
But it suggests that in relation to the fundamental economic pro-
cesses examined here, such differences are better treated as fac-
tors creating variations in the common pattern rather than as the
foundations of entirely different social systems. Greco-Roman as
well as Indo-Muslim civilisation have, in other words, not been
individually reified as absolute, all defining entities, but rather
treated as similar kinds of imperial high culture which developed
in response to a shared condition - a social technology designed
or employed to handlerelated problems arising out of the need to
mobilise the surplus of the peasant masses.7
In Chapter 2, the Mughal parallel suggested that the economic
impact of the Roman Empire should be seen both in terms of elite
7 See Gellner l988b for such an approach, using the term agro-literate society.
293
 EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
building and taxation. When that is done, it becomes clear that
the economic integration created by the formation of empire in
the Mediterranean happened as a consequence of tribute extrac-
tion. It was not a market-driven process, Ricardo-style, basedon a
steadily deepening interregional division of labour. A hypothetical
quantitative model helped to demonstrate that there could hardly
have been room for such a development. The existence of a world
empire, to borrow Wallerstein's term, did not so much prevent
economic integration per se as create a different interregional flow
of the limited agricultural surplus. The function of the market
was here rather that of serving as a transformer of the politically
extracted agricultural surplus. In that role market trade greatly
facilitated the capacity of the political elite to consume the surplus
outside its production area.
The second part ofthe study, "Imperial Bazaar",examinedhow
this different function ofMughal and Roman traders alike affected
the shape of the market world in institutional terms. A key con-
cept was the organisational capacity of merchants. Neither Mughal
nor Roman traders acquired as solid and influential a position in
the political order of their respective societies as early modem
European traders did. The consequences of this were brought out
in Chapter 3 by an analysis of the clash of European and Indian
trading institutions that occurred within the Mughal Empire. The
more limited social power of Indian merchants made for a lesser
ability, on average, to integrate markets. Individualmarkets in the
Mughal Empire were therefore more prone to violent short-term
fluctuations and were characterised by considerable fragmenta-
tion. It was a less predictable trading environment than gradually
began to emerge in the Dutch and later British-led European world-
system. This model wastested, to the extent that waspossible, with
positive results on Romanmaterial, includingprice series from the
province of Egypt. From here the parallel with Mughal Indiawas
expanded to include consideration of the ability of the imperial
apparatus to homogenise coinage, laws and measures. Often the
Roman Empire has been seen as creating an institutionally uni-
fiecl economic space in the lands surrounding the Mediterranean.
The Mughal comparison helped bring out the clear limits to what
294
EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
was achieved by the tributary empires on that account. Though a
homogenising influence, the trading world of the Roman Empire
remained steeped in local and regional practices.
The many local variations and irregularities, however, must not
be mistaken for complete chaos and fragmentation. Instead Mughal
Indiasuggests that the merchants developed a market form with a
particular shapeof its own. It is normally referredto as the bazaar.
But it is not the "castrated" phenomenon, sidelined by modem
developments, of petty retailers we meet as tourists in poor Mid-
die Eastern countries today. These are only the sad remains of a
formerly vibrant commercial universe. The larger bazaars of the
agrarianate world included long-distance trade, wholesalers, bro-
kers, auctions andspecialisedmoney lenders. However, if modem
capitalist markets have been characterised by a drive to integrate
separate markets, stabilise and homogenise trading conditions,
increase security and reduce irregularities of trade, the bazaar oper-
ated at a more modest level of ambition. Its ability to transform the
commercial environment and provide more generally stable condi-
tions was much less. But, as argued by Clifford Geertz, the bazaar
merchant was not reduced thereby to the level of a powerless vic-
tim of irregularity. Instead he developed strategies to manage or
possibly even take advantage of the situation. The bazaar was a
world of many impenetrable, tightly knit groups seeking to profit
from local, most often short-term, monopolies. The much dis-
cussed famine in Antioch during the visit of the emperor Julian
presents us with precisely this form of commercial speculation.
The chapter therefore concluded by suggesting the bazaar as a
general model for the workings of Roman commercial life.
The mechanisms and strategies developed by the bazaar mer-
chants in the face of irregularity were explored further in the two
following chapters. Chapter 4 examined the relationship between
bazaar and state through the process of customs collection. On
the basis of the Mughal comparison it was portrayed as a balance
of power. In spite of the customs collector's proverbial abusive-
ness, the ability of merchants to go elsewhere did impose some
limits on his exactions. Though far from satisfying modem notions
of security this balance of power did provide a fragile order that
295
 EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
enabledtradeto take place. Ifthe imperial state didnot providethe
level ofguaranteesofferedby thebourgeoisstate, the merchantsof
the bazaarcompensated by developing private forms of order and
organisation. This was treated in Chapter 5, whichexplored com-
munity building and social networks among the merchants in the
bazaar. The focus was on communal structures and the household.
By comparing Indian caste with Roman collegia it was suggested
that a significant economic role underlay the often-noticed pre-
dominantly socio-religious character of Roman professional asso-
ciations.Thoseassociationswereimportant in creatingor support-
ing structures of community among the merchants; social norms
were disseminated, privatejudicial arbitration and credit possibly
made available. The household, too, was an important source of
socialsupport. It providedthebasisofcommercialcapitalandfitted
into a hierarchicalpattern of commenda-stylefinancing.However,
by fostering clientelisation, to use Geertz' expression, both sets of
institutions reinforced the fragmented shape of the bazaar. They
benefited insiders while making it very difficult for outsiders to
penetrate markets. This would have created a tendency for goods
to be channelled into compartmentalised and separate circuits of
distribution.
This studyhasaimedtoprovidea differentcontextforourunder-
standing of market trade in the Roman Empire. The prolonged
battle of the modernists and primitivists has tended to descend
into a quarrel over quantities: much or little, important or unim-
portant? This is beside the point. Clearly large quantities of goods
did move about in the Roman Mediterranean and markets served
an important function in the empire. But the aristocratic and trib-
utary order hadvery differentneeds from capitalism. Thoseneeds
fashioned the shape of the world of market trade in a particular
way and promoted distinct commercial patterns which differed
from the evolving capitalism of early modem Europe in signifi-
cant ways. Roman tradewasnot bourgeoisor capitalist. It tookthe
form of market exchange characteristic of the civilised societies
of the agrarian world: the bazaar.
The bazaar, however, does not only represent a particular way of
organising trade. Institutions have been the focus of this study. But
296
EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
the concept ofthe bazaaris farfrom exhaustedby that approach. It
canbedevelopedin otherdirections, too. Inthepopularperception
the bazaar is primarily a world of a bewildering variety of exotic
rarities and enticing luxuries, a universe speakingto and stimulat-
ing the senses through fine, expensive cloths, delicate handcraft
and aromatic, colourful spices. That reputation is well earned and
is closely connected with sumptuous forms of imperial expendi-
ture. The dominance in societies such as the Roman and Mughal
of consumption by the patrimonial households of the landed elite
andthe emperorhadmore than institutional effects on the imperial
bazaar. It also shapedthe world of goods in fundamental ways.
It is often stated that only with the advent of capitalism did
trade across longer distances in ordinary commodities, staple
goods that is, become important;8 in the preceding aristocratic
world, trade was only a question of luxuries. Outside the Euro-
pean Middle Ages, though even here there are grounds for hes-
itation, this view is in need of substantial revision. Both the
Roman Empire and Mughal India knew large flows of staple
goods - not, however, of the generalised production and market-
ing of goods aimed at the expanding mass-market of the middle
classes that came to fruition during the seventeenth and particu-
lariy eighteenth centuries in, for example, Holland and England.9
But the consumption of elite income and imperial tribute did
intensify patterns of urbanisation and increase demand for sta-
pie goods. The Mughal and Roman emperor or grandee were in
public view, their power constantly displayed in lavish expendi-
ture on pageantry, public buildings(temples, markets etc. ), private
"palaces" and large numbers of servants and retainers to provide
Israel 1989attemptstocountertheexclusivefocusonstaplegoodsinmanydescriptions
of Dutch capitalism.
9 Describedby McKendrick,BrewerandPlumb 1982. See KIoft 1996and Morley 2007:
39-4&for a discussion of ancient forms of consumption. Yet, while it is important not to
write ancient luxury consumption off as a trivial economic phenomenon, we still need
concepts to distinguish different cultures of consumption, instead of simply elevating
modernmiddle-classconsumptionto the universal model. Pomeranz2000, chapter3, for
instance, finds greater similarity between early modern Chinese and European patterns of
consumption, with Indiahaving a much steeper hierarchy of consumers. However, with
the consolidation of the Ch'in dynasty, the Chinese pattern becomes more "imperial",
less egalitarian and "bourgeois".
297
 EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
a suitable backdropfor the grand spectacle of aristocratic life. 10If
thevisionofempireclaimedcommandoftheworld'sdiversitythat,
too, included the availability of men and women in abundance -
as servants, craftsmen, entertainers, producers or just audience.
All this increasedthe sizeofurbanpopulations. The top end ofthe
urban hierarchy, in particular, saw the creation of some very big
conurbations swelledby the enormous expenditureofthe emperor
andleadingaristocrats.The luxury ofimperial culturenecessitated
the mobilisation of large urban populations to serve its needs and
hencecreated a demandfor substantialstaple imports to the impe-
rial capitalsandsomeothercitiesthathadoutgrowntheproduction
capacity of their hinterlands. ' '
The idealofuniversalempirewasactedout in thelifestyle ofthe
emperor, closely emulated by the different strata of the landed elites
to thebestoftheirvaryingfinancialcapacities.Aroundtheimperial
court andthe great households of the leadingnobles an extremely
elaborate culture of luxurious consumption evolved. It was based
on abundance and artistic refinement in art, handicrafts and haute
cuisine,notto mentionlove-makinganderoticismaswell asphilo-
sophical reasoning. Of course there were differences in emphasis
between the two courtly cultures. Not every register was played to
the same extent or in the same way. The art of calligraphy, as any
visitor to the Taj Mahal will testify, was far more important to the
Muslim Mughals than to Roman emperors who on the other hand
promoted human sculpture to a much larger extent, to the benefit
oflaterEuropeanmuseums. Butbothcultures setgreatstore by the
availability of the most valuable, prized and exotic luxuries such
as precious stones and metals, expensive colours, rare materials,
See Blake 1991 on Mughal courtly consumption. Friedlander 1921-3 passim pio-
neered the study of Roman styles of consumption, e. g. vol. I: 33-35; vol. 2, chap. xi,
and attempted to sever it from the realm of moralistic condemnation and fantasy. An
important point of his was that modern capitalist consumption had exceeded Roman.
Veyne l976b istheclassictreatment ofpublic munificencein the Hellenistic andRoman
worlds. For modem analyses, see Gamsey 1994, chapters 8-9 (food); Toner 1995
(leisure); Edwards and Woolf 2003 (the various aspects of imperial Rome). Weeber
2003; 2006 is a useful, but mainly descriptive, survey of Roman luxury consumption.
'' SeeHabib& Raychaudhuri[982: 170-171and Richards 19930:28onthemainMughal
cities and strongholds. On Mughal urban hierarchies and revenue flows: Moosvi 1987,
chapters 12-13; Blake [987; Richards 1987 (contributions of Blake and Richards); on
Roman: Woolf 1997.
298
EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
animals and spices frequently imported from afar and even beyond
the confines of the empires. '2 Palmyrene grave goods provide
ample documentation for the import of Chinese silks into the
Roman Empire via India. Spices and incense were other products
which the Romans imported from the East. 13
The notion of the marvellous and the singular was central to this
kind of consumption. Monarchs and aristocrats delighted in the
presentation of wonderful and sensational specimens of animals
and objects. Scattered across the memoirs of the Mughal emperor
Jahangir are many expressions of satisfaction prompted by the
reception of such objects:
ThisdayDilawarKhan'ssoncamefrom Patan. . .andpaidhomage,presenting a
Kachhi horse. It was extremely well formed and easy to ride. No horse this good
had been presented since I came to Gujarat... On this day my son Shahjahan
showed me a pomegranate and a quince he had been brought from Farahabad.
Until now nothing this big had ever been seen. '4
Ostentatiousexpenditureby thearistocratwasmeantto impress
andto dazzlenotonly thecommonfolk butalso, andmore impor-
tantly, his peers. Roman and Mughal nobles joined in competi-
tion over who could stage the most impressive, surprising and
extraordinarydisplays. The fashions of luxury were driven by the
agonistic impulse of the political elite. Occasionally one would
see attempts to curb this competitive urge to consume. Sumptu-
ary laws and moralising discourse, harping on the corruption of
manners or the frenzy of fashion and teaching personal restraint,
served intermittently to reinforce cohesion among the aristocra-
cies and to protect their members from committing financial sui-
cide in a kind of gigantic imperial potlatch. True nobility, that
was the message, depended on more than the mere ability to
spend. '5
Dalby 2000: I aptly notes that "a geography of luxury runs though the literature of
imperial Rome". Unfortunately his book is mainly descriptive rather than analytical, a
catalogue of ancient literary references.
13 Schmidt-Colinet, Stauffer and AI-As'ad 2000.
14 The Jahangirnama, trans. Wheeler M. Thackston, Oxford, 1999, p. 251. Phlegon's
Bookof Man'els (conveniently translated by William Hansen, Exeter, 1996)is a good
example of Greco-Roman fascination with the strange and extraordinary.
15 On these aspects of elite consumption, see Edwards 1993, in particular chapter 5.
299
 EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
Since Sombart wrote his now classic work on Luxury and Cap-
italism, it has become customary to see modem middle-class con-
sumption as emerging out of the tension between the democratic
force of fashion and the rearguard actions of aristocratic groups
trying to limit competition by making key status markers their
legal preserve. As the Middle Ages gave way to the early modem
era in Europe, old styles of aristocratic living became obsolete.
Wherefeudalnobleshadgloriedin the sizeoftheirretinues,early
modem elites placed a growing emphasis on refinement in hous-
ing, dining and courtship. At the end of this process, aristocratic
privilege gave way to bourgeois spending. Consumption became
increasingly shaped by the middle classes and access to the mar-
ket. 16In theoretical terms, this represented a shift in the culture of
consumptionfrom a coupon-to a fashion-regulatedsystem. Access
to prestige goods was not reserved to groups with particular privi-
leges and entitlements (coupons). Instead the ability to follow the
changing dictates of fashion came to define high status. 17
However, there is a danger of imposing too teleological a view
on these developments. While the rule of fashion may in some
contexts serve to break down established hierarchies and widen
participation to include broadergroups, there is nothinginevitable
or automatic aboutthis outcome, not even in theEuropeancontext.
The possession of vast retinues remained central to the Roman and
Mughal nobles as they diversifiedtheir habits of consumption.
The dichotomy of stable coupon and dynamic fashion systems
seems too crude to do justice to the Roman and Mughal experi-
ence. Both imperial societies were far too complex and dynamic
to rely primarily on law and established convention to regulate
This view of the history of European consumption was anticipated by Ferguson in his
Essay on the History of Civil Society of 1767, part sixth, section III (1995: 238-241).
17 Cf. Appadurai 1986: 32. But as his own discussion, pp. 23-41, indicates, the operation
offashiondoesnot necessarilyunderminearistocraticdominance.Thepolardistinction
betweencouponandfashionmay usefully capturethedifferencebetweenmany simple
societies known from anthropology and modem consumption; it is less well suited to
understand differences between complex agrarian, or as I prefer to call them, agrarianate,
societies.
18 Horace, Serm. 1, 100-130, by literary conceit, considered himself fortunate that he was
sufficiently unimportant to be able to travel without a huge retinue. Yet, Horace, the
'simple-Iiving' poet of high equestrian substance, still took his modest meal served
by three(!) slaves. See Blake 1991, chapters 2-4 on the importance of the great noble
households in the Mughal Empire.
300
EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
aristocratic entitlement to prestige goods, though they did that too.
Their aristocraciesmobilisedmarkets successfully to obtain access
to a widened range of luxuries to adorn their dwellings and grace
their ambience. The elaboration of the culture of consumption did
not, in their cases, lead to any notable middle-class challenges
to their hegemony. It is true that Roman satirists never seem to
tire of lampooning the unworthy upstart. Pride of place is taken
by Petronius' famous literary portrait of the stupendously wealthy
freedman Trimalchio who, in spite of his riches, fails miserably
to win respect in the eyes of the aristocratic readership; his social
pretensions are held up to ridicule and terminate in farce. But, as
Paul Veyne argued long ago, the fleeting class of freedmen could
hardly represent a serious social challenge to the Roman elite. 9 It
did,however,providea convenienttargetof scornagainstwhichto
reaffirm elite identity. There is nothing intrinsically "democratic"
about fashion-driven consumerism. The Roman and Mughal ver-
sion served, rather, to consolidate the position of the imperial aris-
tocracies and increasethehierarchicalgradientin society. Adam
Ferguson, one of the Scottish Enlightenment philosophers, under-
stoodclearly thatthe emergingcommercial societyofhis ownday
was something different from the Roman past. "The manners of
the imperial court, and the conduct of succeeding Emperors will
scarcely gaincreditwiththose whoestimateprobabilitiesfrom the
standard of modem times", he commented of Rome.
Here again the bazaar, with its emphasis on the irregular, seems
a useful prism through which to observe the imperial styles of
consumption rather than treating them in an analytical framework
derived from the emerging middle-class consumer revolution of
the seventeenth, eighteenth and nineteenth centuries. One aspect
ofthe repertoire ofgoods in the bazaaris the relatively low degree
of standardisation. But merchants could also turn this characteris-
tie into a virtue by stressing the unique properties and particular
quality of their products. That enabled them to respond to and
19 Veyne 1961. Cf. Nicolet 1980 on the lack of a middle-class "challenge"to the Roman
aristocracy.
20 See Qaisar 1967 for the steep economic hierarchy of the Mughal Empire. Gamsey
1970remains fundamental on the intensification of social hierarchiesunder the Roman
imperial monarchy.
21 Ferguson 1799, vol. 5: 404 (with Pocock 2003: 410-411).
301
 EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
cultivate the demand of the imperial courts and aristocracies. The
ideological principle underlying this form ofconsumption is neatly
captured by C. A. Bayly:
Whereas modern complexity demands the uniformity of Levis and trainers, the
archaic simplicity of everyday life demanded that great men prized difference
in goods .. . and sought to capture their qualities. Modem "positional"goods are
selt-referential to themselves and to the markets that create demand for them; the
charismatic goods of archaic globalization were embedded in ideologies which
transcended them. In one sense archaic lords and rural leaders were collectors
rather than consumers. What they did, however, was more than merely to collect
becausepeople,objects, foods,garmentsandstyles ofdeportmentthusassembled
changed the substance of the collector... [the archaic lord aimed] strategically
to consume diversity. 22
The collection and consumption of exotic rarities and novel
products was, like Pliny's Historia Naturalis, a cultural manifes-
tation of the wide reach of imperial might. 23 As such this activity
formed an important part of the self-fashioning of Roman and
Mughal nobles as gentlemen and as men of power who were
free of ordinary constraints and could realise the full potential
of human life in all its aspects. 24 "Clothes maketh the man", as
the expression goes. Luxury goods were thus imbued with cer-
tain charismatic, mystical or even magical properties; they were
culturally invested with transformative capacity. Expensive spices
andotherrare natural substanceswere regularly used in medicine,
muchofit basedonhumoraltheoriesofbalancingbodilyelements.
Equally, precious goods were used extensively in religious ritual.
The burning of incense, and other fragrant substances, was a staple
of solemn ceremony. 25
While such esoteric goods - "evocative singularities" as
Marshall Sahlinscalledthem in the epigraphto theseepilegomena
22 Bayly 2002:52-53.
23 SeeBeardandHenderson2001:96-105andchapter4 for someobservationson power
and the art of collecting in ancient Rome.
24 See Finley 1985(1, chapter 2 for the relevance of the gentlemanly ideal for ancient
economic history; Blake 1991,chapter5 and Richards 1984on gentlemanly aristocratic
culture in the Mughal Empire.
25 As Miller 1969, chapters 1-2, made clear, much of our evidencefor the ancient spice
trade is derived from medical treatises, e. g. Galen's De Antidosis. Forthe central position
of humoral theories of the body to the consumption of food, drink and medication, see
further Galen, On the Properties of Foodstuff's and On the Natural Faculties II, 8.
302
EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
- facilitated the creation of public authority and the cultivation
of elevated character, they also served as a test of nobility. These
rare, often foreign and always ruinously expensive products were
surrounded by a dangerous aura; they were difficult to manage and
control. Only the truly noble character knew how to master them,
that was the claim. This belief was preached among the elites to
bolster their confidence and preserve their sense of exclusivity.
Therefore the use of prestige goods was enveloped in complex
rituals and civilisational codes. Intricate and many of them tacit,
the rules of decorum were all important: "Let no one rashly claim
as his the art of dining. "26 These words Horace, the protege of
the arbiter of taste at the Augustan court, Maecenas, with double-
edged irony put into the mouth of one of his satirical characters
who is exposed as having failed himselfto grasp the true content
of what he was admonishing. Money was only a prerequisite to
participation in polite society. Exterior possession ofknowledge of
the proper manners and rituals of consumption was a "give-away"
and socially inhibiting in higher circles. Horatian satire merci-
lessly exhibits the failings of the trite big spender who, for all his
display of luxuries, falls far short of the civilisational idiom and
therefore comes across as vulgar and embarrassing. A host who
bores his guests with long lectures on the splendours ofthe dinner
he is serving them, instead of trusting their refinement and abil-
ity to appreciate the delicacies on their own, is trying too hard to
impress.27 The secret to social success wasthe effortless mastery
of the art of consumption which came from an internalisation of
the proper manners, acquired through long exposure. The great
man was no mere commoner who had become rich. He was a con-
noisseursurroundinghimselfwithall thetrappings ofhighculture
andknowinghowto put themto theirproper andbalanceduse, the
true embodiment of their spirit. 28
HOT. Serm II, 4, 35: "nec sibi cenarum quivis temere arroget artem" (trans. Frances
Muecke, Warminster, 1993). Decorum is a key concept in Cicero's De Officiis, a manual
on the proper gentlemanly conduct. An example of a Mughal gentleman's manual is
provided by Husain 1913 (with the analysis of Blake 1991: 130-132).
27 Hor. Serm. II, 8.
28 AsVeblenremarkedinhisTheoryoftheLeisureClass(1899:47):"thegentlemanmust
consume freely", "of the right kind" and"in a seemly manner".
303
 EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
The patterns of consumption developed by the aristocratic lead-
ers were emulated by local elites across the empires and in some
areas trickled even further down the social hierarchy. Trade in fine
food grains and fruits, Kashmir! shawls, Bengali muslins, silks
from Gujarat and other prestige goods catered to the demand of
the households of Mughal noblemen, wmindars, rural leaders and
service gentry of the small towns.29 In the Roman Empire, the
ruins of Pompeii document how the "high" styles of decoration
were emulated by middling urban households, though of course in
a modifiedversion. 3 In the Gaulic provinces Roman-style goods,
in a limited range, also made their way to the more prominent
peasants such as village leaders.3' The dissemination, portrayed
by Paul Zanker, of imagery andstyles related to the imperial court
across Roman provincial societies was perhaps, rather than the
result of propaganda, an expression of the wish by local grandees
to participate in this culture of consumption, acquire the substance
of universal hegemony. 32 As courts and aristocracies set new stan-
dards for prestige objects, the range and diversity of material cul-
ture expanded, as archaeologists will testify, in both empires.
Thus, the imperial bazaar does not only offer a different insti-
tutional context for our understanding of markets, it also opens a
window to a different world of goods, a distinct culture of con-
sumption. Under the Romans, this is what unites the marketing
of large quantities of the mould-made terracotta tableware known
as terra sigillata, one of the most common artefacts of Roman
archaeology, and the imports of spices, cloths and ivory from the
greater region of the Indian Ocean described in the Periplus Maris
Erythraei.33 As a concept, the bazaaris broad enough to accom-
modate both the peddling of small batches of ornamented, red-slip
tablewares by traders travelling with a few pack animals between
the rural fairs and numerous villages of the Roman countryside,
29 Bayly 1983, introduction and chapter l; l986b: 297-302.
30 Wallace-Hadrill 1994, chapter 7. 3I See Woolf 2001 for this suggestion.
3 See Zanker 1988 on dissemination of styles; Morley 2007: 46-49 on the importance of
imitation.
33 As a concept the bazaar allows us to steer clear of some of the modernising connotations
implied by using the modern notion of globalisation - as does Hingley 2005 - to
understand the Roman Empire. There were some significant differences between the
two processes. See Casson 1989 for the modem standardedition of the Periplus.
304
EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
and the considerably fewer merchants traversing the Middle East
and Indian Ocean to bring back each year a fairly limited num-
ber of more valuable cargoes. 34 In the latter case, the increased
demand created by the development of imperial styles ofconsump-
tion marks a conjuncture in the history of old-world long-distance
trade. It is in the Roman period that the pattern familiarfrom later
ages of precious metals flowing eastwards in return for Oriental
spices, incense and cloth really takes shape and matures. From
India as well as other foreign locations, these goods and many
more arrived in the Roman world, particularly through Egypt and
Syria, with Alexandria and the caravan city of Palmyra as impor-
tant commercial hubs. 35
This basic system was to endure for centuries, though the for-
tunes of individual centres fluctuated. By the sixteenth century,
Palmyra had long since slipped back into obscurity and stopped
playing any role in long-distance trade. That function was now
served by Aleppo while Cairo had eclipsed Alexandria as the
main market in Egypt. Since Roman times, this trade had gone
through several ups and downs. But underneath the movement of
ebb and flow, the system had continued to develop and expand,
not least under the aegis of the creation of a vast Muslim ecumene
stretching from North Africa into Central Asia andIndia.3 With
the consolidation of the strong imperial monarchies of the Ming
dynasty in China, the Ottomans in the Mediterranean, the Safavids
in Persia and the Mughals in India, the pace and size of old-world
long-distance trade received new stimulus from increased demand
during the fifteenth and sixteenth centuries. It was to tap into the
wealth of this trade in general and to cut into the profits earned
by the Venetians in their capacity as the leading middlemen in
the Levantine trade in particular that Europeans, spearheaded by
34 The enormous values concentrated in relatively small quantities of Eastern prestige
goods are best documented by the so-called Muziris Papyrus which records a single
cargoworthalmostHS7 million (7 times theminimumfortunerequirementfor a Roman
senator). For edition and translation see Casson 1990 and Rathbone 2001.
35 Raschke 1978 surveys much of the evidence and provides a sceptical antidote to the
most optimistic speculations and wildest guesses hazarded about this emerging trade.
See Tchemia 1992 and 1995; De Romanis 1996; Yong 2001 for a selection of more
recent work on Rome's Oriental trade.
3 Chaudhuri 1985and 1990;Abu-Lughod1989;Curtin 1984.
305
 EPILEGOMENA: TAKING STOCK - THE WORLD OF GOODS
the Portuguese, circumnavigated Africa and stumbled on America.
Neither of these events, as used to be thought, brought down the
old system which, after some initial disruption, continued to func-
tion. 37 But the establishment of commercial empires in the Indian
Ocean, notably by the Portuguese, the Dutch and the English,
combined with the arrival of American silver, did begin to modify
the contours of the pre-existing pattern. From the late seventeenth
century, this process intensified and European traders began to
subject and transform the old system more deeply. 38 That spelled
the end of the "imperial bazaar". The Mughal Empire thus stands
at the end of a historical conjuncture which, irrespective of some
earlier preliminaries, came into its own for the first time under the
Romans. This adds historical depth to our comparison. For these
were two tributary empires whose trade and markets followed
similar patterns; they shared the world of the imperial bazaar.
37 Das Gupta and Pearson 1987; Prakash 1997, introduction. See Periin 1993 for some
further, if perhaps occasionally too grand, observations on this world.
3 Bayly 2004, chapters 1-2 are now fundamental. Steensgaard l973a remains the best
structural analysis, and further 1987. This also expandedthe quantities involved enor-
mously. In Zosimus' account of Alaric's first siege of Rome (V, 41, 4), a ransom was
agreed involving the payment of 3,000 pounds of pepper. By implication this was sup-
posed to represent a very substantial consignment of the spice in late antiquity. By the
end of the seventeenth century, the combined annual imports to Europe of the Dutch
andEnglishEastIndiaCompaniesamountedto some7 million pounds, seeSteensgaard
l99oa: 120. Obviously, these figures are not directly comparable. But there is no way
to estimate total annual Roman pepper imports, so we are left with impressions. They
suggest a change in the orders of magnitude; the Roman ransom would not have made
a huge impression in the eighteenth century. As Gibbon concluded (1993, vol. 3: 277
and chapter 31): "the improvement of trade and navigation has multiplied the quantity
andreduced the price [of pepper]".
306
BIBLIOGRAPHY
Abramenko, A. 1993, Die Municipale Mittelschicht in kaiserz. eitlichen Italien,
Frankfurt.
Abu-Lughod, J. 1989, Before European Hegemony. The World System A. D.
1250-1350, Oxford.
Adams, C. E. P. 1995, "Supplying the Roman Army: 0. Petrie 245", Zeitschrift
fiir Papyrologieund Epigraphik 109, pp. 119-124.
2007, Land Transport in Roman Egypt. A Study of Economics and Adminis-
tration in a Roman Province. Oxford.
Adams, J. 1994, "Trading States, Trading Places: The Role of Patrimonialism
in Early Modem Dutch Development", Comparative Studies in Society and
//toorv36, pp. 3i9-355.
Adanir, F. 1997, "Der Zerfall des Osmanischen Reiches", in A. Demandt (cd. ),
Das Ende der Weltreiche. Von den Persern bis zur Sowjetunion, Munich,
pp. 108-129.
Ain-i Akhari, by Abu'1-Fazl Allami, translated into English by H. Blochmann,
ed. D. C. Phillott, 3 vols., 2nd edn, Calcutta, 1939.
Akerlof, G. 1970, "The Market for 'Lemons'", Quarterly Journalof Economics
84, pp. 488-500.
Alam, M. 1986,The Crisis of Empire in MughalNorth India, NewDelhi.
Alam, M. and Subrahmanyam, S. (eds. ) 1998, The Mughal State 1526-1750,
Delhi.
2000, "Witnessing Transition: Views on the End of the Akbari Dispensation",
in Panikkar,Byres andPatnaik, pp. 104-140.
Alcock, S. 1993, GraeciaCapta. The LandscapesofRomanGreece,Cambridge.
i997a> "Greece: A Landscapeof Resistance?", in Mattingly i997a, pp. 103-
"5.
(ed.) l99?b. The Early Roman Empire in the East, Oxford.
Alcock, S., D'Altroy, T., Morrison, K. D. and Sinopoli, C. M. (eds.) 2001,
Empires. Perspectivesfrom Archaeology and History, Cambridge.
Alfoldy, G. 1977, Konsulat und Senatorenstand unter den Antoninen. Proso-
pographische Untersuchungenzur SenatorischenFiihrungsschicht,Bonn.
Ali, M. A. 1975, "The Passing of Empire: The Mughal Case", Modern Asian
Studies 9, pp. 385-396.
1985, The Apparatus of Empire. Awards of Ranks, Offices, and Titles to the
Mughal Nobility, 1574-1658, Delhi.
307
 BIBLIOGRAPHY
1986-7 "Recent Theories of Eighteenth Century India", Indian Historical
Review 12, 1-2, pp. 102-1 10.
l997> The Mughal Nobility under Aurangzeb, Delhi (1st edn, 1966).
Alston, R. 1995, Soldier andSociety in Roman Egypt, London.
1998, "Trade and the City in Roman Egypt", in Parkins and Smith, pp. 168-
202.
2002, The City in Roman and Byzantine Egypt, London and New York.
Amin, S. 1973, Le Developpement inegal, Paris.
Amouretti, M. C. and Brun, J. P. (eds. ) 1993, La Production du vin etde I'huile en
Mediterranee [Oil and Wine Production in the Mediterranean Area}, Actes
du Symposium International organise par Ie Centre Camille Julian et Ie
Centre Archeologique du Var, Aix-en-Provence et Toulon, 20-22 novembre
1991, Athens and Paris.
Amphores romaines et histoire economique. Dix ans de recerche. Collection de
1'Ecole Fran§aise de Rome 114, Rome 1989.
Anderson, A. C. and Anderson, A. S. (eds. ) 1981, Roman Pottery Research
in Britain and North-West Europe. Papers Presented to Graham Webster,
Vols. 1-2, BAR International Series 123, Oxford.
Anderson, P. 1974, Lineages of the Ahsolutist State, London.
Andre, J. -M. and Baslez, M. F. 1993, Voyager dans I'antiquite, Paris.
Andreau, J. 1974, Les Affaires de Monsieur Jucundus, Rome.
1980, "Reponse a Yvon Thebert", Annales ESC, pp. 912-919.
1987, La Vie financiere dans Ie monde romain. Les metiers de manieurs
d'argent (IVe siecle av. J. -C. -IlIe siecle ap. J. -C. ), Rome.
1994a, "La cite romaine dans ses rapports a 1'echange et au monde de
1'echange", in Andreau, Briant, Descat et al., pp. 83-98.
l994b, "L'ltalie imperiale et les provinces; desequilibre des echanges et
flux monetaires", in L'halie d'Auguste a Diocletien. Collection de 1'Ecole
Frangaisede Rome 198, Rome, pp. 175-203.
I995. "Vingt ans apres L'Economie Antique de Moses I. Finley", Annales.
Histoire, Sciences Sociales 50, pp. 947-960.
;997> "Deux etudes sur les prix a Rome: les "mercuriales" et Ie taux d'interet",
in Andreau, Briant, Descat, et al., pp. 105-120.
1999- Banking and Business in the RomanWorld, Cambridge.
2003, "Lescommer^ants, 1'eliteet la politiqueromainea la fin delarepublique
(IIIe-Ier siecles av. J. C. )", in Zaccagnini, pp. 217-243.
20o6, "Roman Law in relation to Banking and Business: A Few Cases", in
Bang, Ikeguchi andZiche, pp. 201-214.
Andreau, J., Briant, P., Descat, R. et at. (eds. ) 1994, Economie antique.
Les echanges dans I'Antiquite: Ie role de I'Etat, Saint-Bertrand-de-
Comminges.
Andreau, J., Briant, P., Descat, R. et al. (eds. ) 1997, Economie Antique. Prix
et formation des prix dans les economies antiques, Saint-Bertrand-de-
Comminges.
308
BIBLIOGRAPHY
Andreau, J., France. J. and Pittia, S. (eds. ) 2004, Mentalites et choix economiques
des Romains, Bordeaux.
Anctreau, J. and Hartog, F. (eds. ) 1991, La cittu antica? La Cite antique. Opus
6-8, 1987-9.
Ankum, H. 1972, "Quelques problemes concemant les ventes aux encheres en
droit Remain classique", Studi in Onore de Gaetano Scherillo l, Milan,
PP. 377-394.
Aperghis, M. 2001, "Population- Production- Taxation- Coinage. A Model of
the Seleukid Economy", in Archibald et at., pp. 69-102.
2004, The Seleukid Royal Economy. The Finances and Financial Administra-
t'wn of the Seleukid Empire, Cambridge.
Appadurai, A. (ed. ) 1986, The Social Life of Things. Commodities in Cultural
Perspective, Cambridge.
Arangio-Ruiz, V. 1950, La societa in diritto romano, Naples.
Arasaratnam,S. 1986,Merchants, CompaniesandCommerceonthe Coromandel
Coast 1650-1740, Delhi.
1990. "Recent Trends in the Historiography of the Indian Ocean", Journal of
World History 2, pp. 225-248.
Archibald, Z. H. et at. (eds. ) 2001, Hellenistic Economies, London.
Atsma, H. and Burguiere, A. (eds. ) 1990, Marc Block aujourd'hui: Histoire
comparee et sciences societies, Paris.
Aubert, J. -J. 1994, Business Managers in Ancient Rome. A Social and Economic
Study, Leiden.
(ed. ) 2003, Taches publiques et entreprise privee dans Ie monde romain,
Geneva.
2004, "De 1'usage de 1'ecriture dans la gestion d'entreprise a 1'epoque
romaine", in Andreau, France and Pittia, pp. [27-148.
Aubert, J. -J. and Sirks, B. (eds. ) 2002, Speculum luris. Roman Law as a Reflection
of Social and Economic Life in Antiquity, Ann Arbor, MI.
Ausbiittel, F.M. 1982,Untersuchungenw denVereinenimWestendesrmnischen
Reiches. Kallmiinz.
Autour de Coptos. Actes du colloque organise au Musee des Beaux-Arts de
Lyon (17-18 mars 2000). Topoi. Orient-Occident 3, Supplement, Lyon,
2002.
Aymard, M. (ed. ) 1982, Dutch Capitalism and World Capitalism, Cambridge.
1990, "Histoire et comparaison", in Atsma and Burguiere, pp. 271-
278.
Baechler, J., Hall, J. and Mann, M. (eds. ) 1988, Europe andthe Rise of Capitalism,
Cambridge.
Bagnall, R. S. 1993, Egypt in Late Antiquity, Princeton, NJ.
2000, "Governmental Roles in the Economy of Late Antiquity", in Lo Cascio
and Rathbone, pp. 86-91.
Bagnall, R. S. and Frier, B. W. 1994, The Demography of Roman Egypt, Cam-
bridge.
309
 BIBLIOGRAPHY
Bagnall, R. S., Helms, C. and Verhoogt,A. 2000, Documentsfrom Berenike,vol.
I: Greek Ostrakafrom the 1996-1998 Seasons, Papyrologica Bruxellensia
31, Brussels.
Bairoch, Paul 1993, Economics andWorld History: Myths and Paradoxes,New
York.
Bajekal, M. 1990,"TheStateandtheRural GrainMarket in EighteenthCentury
Eastern Rajasthan", in Subrahmanyam 19903, pp. 90-120.
Balasz E., 1964, Chinese Civilization and Bureaucracy. Variations on a Theme,
New Haven, CT.
Banaji, J. 2001, Agrarian Change in Late Antiquity, Oxford.
Banerjee, K. K. 1990, "GrainTraders and the East IndiaCompany: Patnaand its
Hinterland in the Late 18th and Early 19th Centuries", in Subrahmanyam
i99oa, pp. 163-189.
Bang, P. F. 19970, Antiquity between "Primitivism" and "Modemism", Ar-
bejdspapirerCenterforKulturforskning,AarhusUniversitet, no. 57,Aarhus,
www. hum. au. dk/ckulturf/pages/publications/pfb/antiquity. htm.
[99?b, "Aristokrati og handel i status-modellen - et clientela-perspektiv",
Historisk Tidssknft 97, i, pp. 1-28.
2002a,"RomansandMughals- EconomicIntegrationin a Tributary Empire",
in De Blois and Rich, pp. 1-25.
2002b, "Trade and Industry", Classical Review, n. s., 52, I, pp. 97-100.
2004, "The Mediterranean: A Corrupting Sea? A Review-essay on Ecology
and History, Anthropology and Synthesis", Ancient West and East 3, 2, pp.
385-399.
2006, "Imperial Bazaar: Towards a Comparative Understanding of Markets in
the Roman Empire", in Bang, Ikeguchi and Ziche, pp. 51-88.
2007, Trade and Empire - In Search of OrganisingConcepts for the Roman
Economy", Past and Present 195, pp. 3-54.
Bang, P. P. and Bayly, C. A. (eds. ) 2003, "Tributary Empires in History: Compar-
ative Perspectives from Antiquity to the Late Medieval", Medieval History
Journal 6, 2, special issue.
Bang, P. P., Ikeguchi, M. and Ziche, H. G. (eds. ) 2006, Ancient Economies,
Modern Methodologies. Archaeology. Comparative History, Models and
Institutions, Bari.
Bang, P. F. et al. (eds. ) 1999, Agrarimperier mellem Marked og Tribut. Den Jyske
Historiker 86/87, Aarhus.
Barbour, V. 1966, Capitalism in Amsterdam in the ijth Century, Ann Arbor, MI.
Barker, G. and Mattingly, D. (eds. ) 1996, Farming the Desert: The
UNESCO Libyan Valleys Archaeological Survey, vols. 1-2, Tripoli and
London.
Bay, A. 1972, Den Romerske Finansadministrationfra Augustus til Hadrian,
vols. 1-2, Horsens.
Bayly, C. A. 1983, Rulers, Townsmen and Bazaars. North Indian Society in the
Age of British Expansion7770-7870, Cambridge.
3io
BIBLIOGRAPHY
19860, "The Middle East and Asia during the Age of Revolutions, 1760-1830",
Itinerario 10, 2, pp. 69-84.
i986b. "The Origins of Swadeshi (Home Industry): Cloth and Indian Society,
1700-1930", in Appadurai, pp. 285-321.
1988, Indian Society and the Making of the British Empire, The New Cam-
bridge History of India 2, I, Cambridge.
1989, Imperial Meridian, Harlow.
2000, "South Asia and the 'Great Divergence'", Itinerario 24, pp. 83-103.
2002. "'Archaic' and 'Modem' Globalization in the Eurasian and African
Arena, c. 1750-1850", in A. G. Hopkins (ed. ), Globaliwtion in World
History, London, pp. 47-73.
2004, The Birth of the Modem World 1780-1914. Global Connections and
Comparisons, Maiden, MA and Oxford.
Bayly, C. A. and Bayly, S. 1988, "Eighteenth-Century State Forms and the
Economy", in Dewey, pp. 66-90.
Bayly, C. A. andSubrahmanyam,S. 1990, "PortfolioCapitalists andthePolitical
Economy of Early Modern India", in Subrahmanyam i990a, pp. 242-265.
Bayly, S. 1999, Caste, Societyand Politics in Indiafrom the EighteenthCentury
to the Modern Age, The New Cambridge History of India 4, 3, Cambridge.
Bean, G. E. 1954, "Notes and Inscriptions from Caunus", Journal of Hellenic
Studies 74, pp. 85-110.
Beard. M. and Henderson, J. 2001, Classical Art. From Greece to Rome, Oxford.
Bekker-Nielsen, T. 2004, The Roads of Ancient Cyprus, Copenhagen.
Beloch, K. -J. 1886, Die Bevolkerung der griechisch-romischen Welt, Leipzig.
Berktay, H. 1991, "Three Empires and the Societies They Governed: Iran, India
and the Ottoman Empire", Journal of Peasant Studies 18, 3/4, pp. 242-263.
Bharadwaj, K. 1985, "A View on Commercialisation in Indian Agriculture and
the Deve\opmentofC&p\talism", Journal of Peasant Studies 12, 1, pp. 7-25.
Bitterli, U. 1991, Die "Wilden" unddie "Zivilisierten", Munich.
Bj0rkelo, A. 1999, "Noen karakteristiske trekk ved 0konomisk virksomhet i Det
osmanske riket", in Bang et at., pp. 121-143.
Blake, S. P. 1979, "The Patrimonial Bureaucratic Empire of the Mughals", Jour-
nal of Asian Studies 39, pp. 77-94.
1987, "The Urban Economy in Pre-Modern Muslim India: Shahjahanabad,
1639-1739"; Modern Asian Studies 21, 3, pp. 447-471.
1991; Shajahanabad. The Sovereign City in Mughal India, Cambridge.
BlazquezMartfnez, J. M. and Remesal Rodnguez,J. (eds. ) 1999-2001,Estudios
sobre el Monte Testaccio (Roma), vols. 1-2, Barcelona.
Bloch, M. 1934, "Une etude regionale: geographie ou histoire?", Annales
d'Histoire Economiqueet Sociale6, pp. 81-85.
1967, Land and Work in Medieval Europe. Selected Papers by Marc Block,
trans. J. E. Anderson, Berkeley and Los Angeles.
Blois, L. de and Rich, J. (eds. ) 2002. The Transformation of Economic Life
under the Roman Empire. Proceedings of the Second Workshop of the
3"
 BIBLIOGRAPHY
International Network Impact of Empire (Roman Empire, c. 200 B. C. -A. D.
476), Amsterdam.
Bogaert, R. 2000, "Les operations des banques de I'Egypte romaine", Ancient
Society 30, pp. 135-269.
Boserup, E. 1965, The Conditions of Agricultural Growth. The Economics of
Agrarian Change under Population Pressure, London.
1981, Population and Technology, London.
Bounegru, 0. 2006, Trafiquants et navigateurs sur Ie Bas Danube etdans Ie Font
Gauche a I'epoque romaine, Wiesbaden.
Bradford Welles, C., Fink, R. 0. and Gilliam, J. F. (eds. ) 1959, The Excavations
at Dura-Europos final report, 5, pt. I: The Parchments and Papyri, New
Haven, CT.
Brady, T. A., Jr. 1991, "The Rise of Merchant Empires, 1400-1700: A European
Counterpoint", in Tracy, pp. I 17-160.
Braudel, F. 1972, The Mediterranean and the Mediterranean World in the Age
of Philip II, vol. I, London.
1981, The Structures of Everyday Life. Civilization and Capitalism, vol. I,
trans. S. Reynolds, London.
1982, The Wheels of Commerce, Civilization and Capitalism, vol. 2, trans. S.
Reynolds, London.
1984, The Perspective of the World, Civilization and Capitalism, vol. 3, trans.
S. Reynolds, London.
Braudel, F. and Spooner, F. 1967, "Prices in Europefrom 1450to 1750", in E. E.
Rich and C. H. Wilson (eds. ), The Cambridge Economic History of Europe,
vol. 4, chapter 7, pp. 374-486.
Braund, D. 1993, "Piracy under the Principate and the Ideology of Imperial
Eradication", in Rich and Shipley, pp. 195-212.
Breckenridge, C. A. and Veer, P. van der (eds. ) 1993, Orientalism and
the Postcolonial Predicament. Perspectives on South Asia, Philadelphia,
PA.
Bresson. A. 2000. La Cite marchande, Bordeaux.
2003, "Merchants and Politics in Ancient Greece: Social and Economic
Aspects", in Zaccagnini, pp. 139-164.
Brimnes, N. 1997, "Traditionaliseringen af Indien", Den Jyske Historiker 77/78,
pp. 13-29.
Bruhns, H. 1985, "De WemerSombart a Max Weber et Moses Finley: LaTypolo-
gie de la ville de consommation", in Leveau, pp. 255-274.
Bmhns, H. and Nippel, W. (eds. ) 2000, Max Weber und die Stadt im Kulturver-
gleich, Gottingen.
Brunt, P. A. 1990, Roman Imperial Themes, Oxford.
Biicher, K. 1911, Die Entstehung der Volkswirtschaft, 8th edn, Tiibingen (first
published 1893).
1922, "Zur griechischen Wirtschaftsgeschichte", in K. Biicher, Beitrage 7,ur
Wirtschaftsgeschichte, Tubingen, pp. 1-97.
312
BIBLIOGRAPHY
Buckland, W. W. 1908, The Roman Law of Slavery. The Condition of the Slave
in Private L(IWfrom Augustus to Justinian,Cambridge.
Barge, A. 1987, "Fiktion und Wirklichkeit: Soziale und rechtliche Strukturen
des romischen Bankwesens", Zeitschrift der Savigny-Stiftung fur Rechts-
geschichte. Romanistische Ahteilung 104, pp. 465-558.
Burke, P. 1988, "Republicsof Merchants in Early Modem Europe", in Baechler,
Hall and Mann, pp. 220-233.
Burton, G. P. 1975, "Proconsuls, Assizes and the Administration of Justice under
the Empire", Journal of Roman Studies 65, pp. 92-106.
Butcher, K. 1988, Roman Provincial Coins. An Introduction to the "Greek Impe-
ricils". London.
Byres, T. J. 2000, "Pre-Capitalist Economic Formations and Differential Agricul-
tural Productivity: A Tentative Hypothesis", in Panikkar, Byres andPatnaik,
pp. 239-275.
Byres, T. J. and Mukhia, H. (eds. ) 1985, Feudalism and Non-Eiiropean Societies,
London.
Cagnat, R. 1882, Elude historique sur les impots indirects chez les Romains
jusqu'aux invasions des barbares d'apres les documents litteraires et
epigraphiques, Paris.
Calder. W. M. and Demandt, A. (eds. ) 1990, Eduard Meyer. Leben und Leistung
eines universal Historiker, Mnemosyne \ 12, Supplementum, Amsterdam.
Camodeca, G. 1999, Tabulae Pompeianae Sulpiciorum (TPSulp. ); Ediz. ione crit-
ica dell'archivio puteolano del Sulpicii, 2 vols., Rome.
2003, "II credito negli archivi campani: il caso di Puteoli e di Herculaneum ,
in Lo Cascio, pp. 69-98.
CapogrossiColognesi,L. 1990, Economicantichee capitalismomoderno, Bari.
1995. "The Limits of the Ancient City and the Evolution of the Medieval
City in the Thought of Max Weber", in Cornell and Lomas, pp. 27-
37.
Carandini,A. 1981, "Sviluppo e crisi delle manifatturerurali e urbane", in Giar-
dina and Schiavone, vol. 2, pp. 249-260.
i983a, "Columella's Vineyard and the Rationality of the Roman Economy",
Opus 2, pp. 177-204.
i983b, "Pottery and the African Economy", in Garnsey, Hopkins and
Whittaker, pp. 145-162.
1988, Schiavi in Italia. Gli strumenti pensanti del Romanifra tarda repubblica
e media impero, Rome.
1989, "Italian Wine and African Oil: Commerce in a World Empire", in Klaus
Randsborg (ed. ), The Birth of Europe, Analecta Romana Institiiti Daniel,
Supplementum 16, Rome.
Cartledge,P.,Cohen,E.E.andFoxhall,L. (eds. ) 2002,Money, LabourandLand.
Approaches to the Economies ofAncient Greece, London.
Casson,L. 1968, Review of Rouge 1966, American JournalofPhilology 89, pp.
359-364.
313
 BIBLIOGRAPHY
1974' Travel in the Ancient World, London.
1984, Ancient Trade and Society, Detroit, MI.
1986, "New Light on Maritime Loans: P. Vinclob G. 19792 (=SB VI 9571)",
in R. S. Bagnall and W. V. Harris (eds. ), Studies in Roman Law in Memory
of A. Arthur Schiller, Leiden, pp. 11-17.
1989,The PeriplusMansErythraeie, ed., trans. andcommentaTybyL. Casson,
Princeton, NJ.
1990 "New Light on Maritime Loans: P. Vindob. G 40822", Zeitschnft fiir
Papyrologie und Epigraphik, 84, pp. 195-206.
Cerati,A. 1975,Caractereannonaireetassiette deI'impotfonderaubas-empire,
Paris.
Chastagnol, A. 1992, Le Senat romain a I'epoque imperiale. Recherches sur la
composition de I'assemblee et Ie statut de ses memhres, Paris.
Chatterjee, K. 1996, Merchants, Politics and Society in Early Modern India,
Leiden.
Chaudhuri, K. N. 1978, "Some Reflections on Town and Country in Mughal
India , ModernAsian Studies 12, pp. 77-96.
l979> "Markets and Traders in India during the Seventeenth and Eighteenth
Centuries", in K. N. ChaudhuriandC. Dewey(eds. ), Economy andSociety,
New Delhi, pp. 143-162.
1981, "The World System East of Longitude 20: The European Role in Asia
1500-1750", Review 5. 2, pp. 219-246.
1985,Trade andCivilizationin the IndianOcean. An Economic History from
the Rise of Islam to 1750, Cambridge.
1990, Asia Before Europe. Economy and Civilisation of the Indian Ocean
from the Rise of Islamto 1750, Cambridge.
Chaudhuri, S. 1983, "The Surat Crisis of 1669: A Case Study of Mercan-
tile Protest in Medieval India!", Calcutta Historical Journal 5, pp. 129-
146.
Chayanov, A. V. 1986, On the Theory of Peasant Economy, with a foreword by
Teodor Shanin, Madison, WI.
Chevalier, R. 1988,Voyages et deplacements dansI'empire romain, Paris.
Christensen, A. E. 1941, Dutch Trade to the Baltic about 1600. Studies in the
SoundToll Register and Dutch Shipping Records, Copenhagen.
Christiansen, E. 1982, "Marx, Slaver og Antikken", Historisk Tidsskrift 82, I,
pp. 63-94.
1988, The Roman Coins of Alexandria, Aarhus.
1995' A History of Rome, Aarhus.
2004, Coinage in Roman Egypt. The Hoard Evidence, Aarhus.
Cipolla, C. M. (ed. ) 1970, The Economic Decline of Empires, London.
(^izak^a, M. 1985, "Incorporation of the Middle East into the European World-
Economy", Review 8, pp. 353-377.
dark, C. and Haswell, M. 1970, The Economics of Subsistence Agriculture, 4th
edn. London.
314
BIBLIOGRAPHY
Clemente, G. 1972,"IIpatronatonei collegiadell'imperoromano", StiidiClassici
e Orientali 21, pp. 142-229.
Coarelli, F. 1988, // Foro Boario, Rome.
Colin, X. 2000, "Commer^ants itinerants et marchands sedentaires dans
1'Occident romain", in Lo Cascio 2000a, pp. 149-161.
Corbier, M. 1991, "City, Territory and Taxation", in Rich and Wallace-Hadrill,
pp. 2i1-239.
Cornell, T. and Lomas, K. (eds. ) 1995, Urban Society in Roman Italy, London.
Cotter, W. 1996, "The Collegia andRomanLaw: StateRestrictions on Voluntary
Associations, 64 BCE-200CE", in Kloppenborg andWilson, pp. 74-89.
Cottier, M. 2003, "Le ferme des douanes en Orient et la lex portorii Asiae", in
Aubert, pp. 215-228.
Cotton, H. 1993, "The Guardianshipof JesusSon ofBabatha:Roman and Local
Law in the Province of Arabia", Journalof Roman Studies 83, pp. 94-108.
Coulton, J. J. 1987, "Opramoas and the Anonymous Benefactor", Journal of
Hellenic Studies 107, pp. 171-178.
Crawford, D. J. 1971, Kerkeosiris. An EgyptianVillage in the Ptolemaic Period,
Cambridge.
Crawford, M. (ed. ) 19860, L'impero romano e Ie strutture economiche e sociali
delle province, Como.
i986b, "The Monetary System of the Roman Empire", in Crawford 19860,
pp. 61-70.
Crone, P. 1989, Pre-industrial Societies, Oxford.
Crook, J. A. 1967, Law andLife of Rome, 90 B. C. -A. D. 212, Ithaca,NY.
1995; LegalAdvocacy in the RomanWorld, Ithaca,NY.
Crow, B. 1999,"ResearchingtheMarketSystemin Bangladesh",in Harris-White
l999at PP-1>5-I5°-
Curtin, P. D. 1984, Cross-CulturalTrade in WorldHistory, Cambridge.
Curtis, R. I. 1991,GarumandSalsamenta. ProductionandCommerce in Materia
Medic a, Leiden.
Dalby, A. 2000, Empire of Pleasures, London.
D'Arms, J. H. 1977, "M. I. Rostovtzeffand M. I. Finley: The Status of Traders
in the Roman World", in D'Arms and Eadie (eds. ), Ancient and Modern.
Essays in Honor of G. E. Else, Ann Arbor, MI, pp. 159-179.
1981, Commerce and Social Standing in Ancient Rome, Cambridge, MA.
D'Arms, J. H. and Kopff, E. C. (eds. ) 1980,The Seaborne Commerce of Ancient
Rome. Studies in Archaeology and History, Memoirs of the American
Academy in Rome 36, Rome.
Das Gupta, A. 1979, Indian Merchants and the Decline ofSurat, c. 1700-1750,
Wiesbaden.
1987, "India and the Indian Ocean in the Eighteenth Century", in Das Gupta
and Pearson, pp. 131-161.
1991, "Changing Faces of the Maritime Merchant", in Ptak and Rothennund,
pp. 352-362.
315
 BIBLIOGRAPHY
1998, "Trade and Politics in EighteenthCentury India", in Alam and Subrah-
manyam, pp. 361-397.
Das Gupta, A. and Pearson, M. N. (eds. ) 1987, India and the Indian Ocean,
/500-/Soo, Calcutta.
Datta, R. 2000, Society, Economy and the Market. Commercialisation in Rural
Bengal, [760-1800, New Delhi.
Daube, D. 1980, "Jewish Law in the Hellenistic World", in B. S. Jackson (ed. ),
JewishLaw in LegalHistory andthe ModernWorld, Leiden, pp. 45-60.
Davies, J. K. 1998, "Ancient Economies: Models and Muddles", in Parkins and
Smith, pp. 225-256.
De Robertis, F. M. 1971, Storia delle corporazioni e del regime associativo nel
mondo romano. vols. 1-2. Bari.
De Romanis, F. 1996, Cassia, cinnamono, ossidiana. Uomini e merci tra Oceano
indiano e Mediterrcmeo. Rome.
l998>"Commercio, Metrologia,Fiscalita",Melangesde I'Ecole Franfaisede
Rome. Antiquite 110, pp. 11-60.
De Salvo, L. 1992, Economia privata e puhblici serviz. i nell'impero romano. I
corpora naviculariorum, Messina.
De Souza, P. 1999, Piracy in the Graeco-Roman World, Cambridge.
Demougin, S. 1988, L'ordre equestre sous les Julio-Claudiens, Rome.
Demougin, S. and Devijver, H. (eds. ) 1999, L'ordre equestre. Histoire d'une
arisfocratie (lie siecle av. J. C -IIIe siecle ap. J. C. ), Collection de I'Ecole
Fran^aise de Rome 257, Rome.
Dewey, C. 1988, Arrested Development. The Historical Dimension, Riverdale,
MD.
Dietz,S.,LadjimiSebai,L. andBen Hassen,H. (eds. ) 1995,AfricaProconsularis.
Regional Studies in the Segermes Valley of Northern Tunisia, vols. 1-2,
Aarhus.
Dirlmeier, U. 1987, "Mittelalterliche Zoll- und Stapelrechte als Handelshemm-
nisse , in H. Pohl. (ed. ), Die Auswirkungen von Zollen und anderen Haii-
delshemmnissen auf Wirtschaft und Gesellschqft vom Mittelalter bis zur
Gegenwart. Vierteljahrschrift fiir Sozial- und Wirtschaftsgeschichte, suppl.
8o, pp. 19-39.
Dirven, L. 1999, The Palmyrenes of Dura-Europos. A Study of Religious Inter-
action in Roman Syria, Leiden.
Disney, A. (ed. ) 1995, Historiography of Europeans in Africa and Asia 1450-
i8oo. An Expanding World, vol. 4, Aldershot.
Dmitriev, S. 2005, City Government in Hellenistic and Roman Asia Minor,
Oxford.
Domergue, C. 1994, "L'etat romain et Ie commerce des metaux a la fin de la
Republique et sous Ie Haut-Empire", in Andreau, Briant and Descat et a!.,
pp. 99-113.
1998, "A View ofBaetica's External Commerce in the 1st c. A. D. based on its
Trade in Metals", in Keay, pp. 201-215.
3l6
BIBLIOGRAPHY
Domergue, C. and Liou, B. 1997. "L'apparition de nonnes dans Ie commerce
maritime romain", Pallas 46, pp. 11-30.
Dreher, M. 1997, "Das Monumentiim Ephesenumund das romische Zollwesen ,
Milnstersche Beitrdge wr antiken Handelsgeschichte 16, 2, pp. 79-96.
Drexhage, H. J. 1988, "... Scimus quam varia sint pretia rerum per singulas
civitates regionesque... Zu den Preisvariationen im romischen Agypten ,
Miinstersche Beitrdge zurAntiken Handelsgeschichte7, 2, pp. l-lI.
1991. Preise, Mieten/Pachten, Kosten und Ldhne irn rdmischen Agypten, St
Katharinen.
i994> "Einfliisse des Zollwesens auf den Warenverkehr im romischen Reich
- handelshemmend oder handelstordernd?", Miinstersche Beitrdge wr
antiken Handelsgeschichte 13, 2, pp. 1-15.
Drexhage,H.-J.,Konen,H. andRuffing,K. 2002, DieWirtschaftdes RSmischen
Reiches (i. -3. Jahrhundert), Berlin.
Drexhage, R. 1988, Untersuchungen wm rdmischen Osthandel, Bonn.
Dreyer, E. L. 2007, Zheng He. China and the Oceans in the Early Ming Dynasty,
1405-1433, New York.
Drinkwater, J. F. 2001, "The Gallo-Roman Woollen Industry and the
Great Debate: The Igel Column Revisited", in Mattingly and Salmon,
pp. 297-308.
Droysen, J. G. 1977, Historik, ed. Peter Leyh, vol. I, Stuttgart-Bad Cannstatt.
Dumont, L. 1980, Homo Hierarchicus. The Caste System and Its Implications,
rev. English, edn, Chicago.
Duncan-Jones,R. 1976, "The Price of Wheat in Roman Egypt underthe Princi-
pate", Chiron6, pp. 241-262.
l979> "Variationin EgyptianGrain Mesure", Chiron 9, pp. 347-375-
1982,The Economy of the RomanEmpire. Quantitative Studies, Cambridge.
1990, Structure and Scale in the Roman Economy, Cambridge.
l994< Money andGovernment in the Roman Empire, Cambridge.
2003, "Roman Coin circulation and the Cities of Vesuvius", in Lo Cascio
2003b, pp. 161-180.
Duthoy, R. 1976, "Recherches sur la repartition geographique et chronologique
des termes sevir augustalis, augustalis et sevir dans 1'empire romain",
Epigraphische Studien n, pp. 143-214.
Eck, W. (ed. ) 1999, Lokale Autonomie und romische Ordnungsmacht in den
kaiserzeitlichen Provinzen vom l. bis 3. Jahrhundert, Munich.
2000, "The Growth of Administrative Posts", in The Cambridge Ancient His-
tory, vol. XI, 2nd edn, Cambridge, pp. 238-265.
Edhem, E., Goffman, D. and Masters, B. 1999, The Ottoman City between East
andWest. Aleppo, Ismir andIstanbul, New York.
Edwards, C. 1993, The Politics of Immorality in Ancient Rome, Cambridge.
Edwards, C. and Woolf, G. (eds. ) 2003, Rome the Cosmopolis, Cambridge.
Hiring, J. and Lund, J. (eds. ) 2004, Transport Amphorae andTrade in the Eastern
Mediterranean. Acts ofthe InternationalColloquiumat the DanishInstitute
317
 BIBLIOGRAPHY
at Athens, September 26-29, 2002. Monographs of the Danish Institute at
Athens 5, Athens.
Eisenstadt, S. 1963, The Political Systems of Empires, London.
Eisenstadt, S. and Schluchter, W. (eds. ) 1998, Early Modernities, Daedalus 127,
3, special issue, Cambridge, MA.
Elvin, M. 1973, The Pattern of the Chinese Past, Stanford, CA.
Empereur, J. -Y. (ed. ) 1999, Commerce et arfisanat dans I'Alexandrie
hellenistique et romaine. Actes du colloque d'Athenes, organise par Ie
CNRS, Ie laboratoire de Ceramologie de Lyon et I'Ecole Franfaise
d'Athenes, 11-12 decembre 1988, Athens.
Empereur, J. -Y. and Garlan, Y. (eds. ) 1986, Recherches sur les amphores grec-
ques, Paris.
Engelmann, H. and Knibbe, D. 1989, "Das Zollgesetz der Provinz Asia. Eine
neue Inschrift aus Ephesos", Epigraphica Anatolica 14, pp. 1-206.
Engels, D. 1990, Roman Corinth, Chicago.
Epigrafia deltaproduzwne e delta distribuwme. Collection de 1'Ecole Fran^aise
de Rome 193, Rome, 1994.
Epigrafia e online senatorw=Atti del Colloquio Internaz. ionale AIEGL su Epi-
grafiae Ordine Senatorio(1981: Roma), Tituli 4-5, Rome 1982.
Erdkamp, P. 2001, "Beyond the Limits of the 'Consumer City'. A Model
of the Urban and Rural Economy in the Roman World", Historia 50,
pp. 332-356.
2002, "A Starving Mob has No Respect: Urban Markets andFood Riots in the
Roman World, 100 B. C.-A. D. 400", in de Blois and Rich, pp. 93-115.
2005,The Roman Grain Market, Cambridge.
Erslev, K. 1911, Historieskrivning. Grundlinier til nogle kapitler af historiens
Theori, Copenhagen.
Etienne,R. 1952,"Romeeut-elle unepolitiquedouaniere?",Annales ESC7, pp.
371-377.
Etienne, R. and Mayet, F. 1998, "Les 'mercatores' de saumure hispanique",
MelangesdeI'EcoleFrangaisede Rome. Antiquite 110,pp. 147-165.
Faroqhi, S. igS^, Towns and Townsmen of Ottoman Aizatolia. Trade, Crafts and
Food Production in an Urban Setting, 1520-1650, Cambridge.
i99i> "In Search of Ottoman History", Journal of Peasant Studies 18, 3/4,
pp. 211-241.
Pawaz, L. and Bayly, C. A. (eds. ) 2002, Modernity and Culture. From the
Mediterranean to the Indian Ocean. New York.
Ferguson,A. 1995,AnEssayontheHistory ofCivilSociety,ed.F.Oz-Salzberger,
Cambridge (first published 1767).
l799> The History of the Progress andTermination of the Roman Republic, a
new edn, revised and corrected, 5 vols., Edinburgh.
Fikhman, I. F. 1991-2, "State and Prices in Byzantine Egypt", Scripta Classica
Israelica 11, pp. 139-148.
Finkielkraut, A. 1988,The Undoing ofThought, London.
3i8
BIBLIOGRAPHY
Finlay, R. 2000, "China, the West and World History in Joseph Needham's
Science and Civilisation in China", Journal of World History 11, 2,
pp. 265-303.
Fintey, M. I. 1954, The World of Odysseus, New York.
1965, "Technical Innovation and Economic Progress in the Ancient World",
Economic Histoiy Review, 2nd ser., l8, pp. 29-45.
1970. "Aristotle and Economic Analysis", Past and Present 47, pp. 3-25.
1975 > The Use and Abuse of History, London.
(ed. ) 1976, Studies in Roman Property, Cambridge.
1977. "The Ancient City: From Fustel de Coulanges to Max Weber and
Beyond", Comparative Studies in Society and History 19, pp. 305-327.
1980, Ancient Slaveiy and Modern Ideology, London.
1983, Politics in the Ancient World, Cambridge.
l9§5a. The Ancient Economy, 2nd edn, London.
i985b, Ancient History. Evidence and Models, London.
Fischer, W., Mclnnis, R. M. and Schneider, J. (eds. ) 1986, The Emergence of a
World Economy 1500-1914, Papers of the IXth International Congress of
Economic History, 2 vols., Wiesbaden.
Foraboschi,D. 1994,"Economieplurali e interdipendenze",in L'ltalie d'Auguste
d Diocletien, Collection de 1'Ecole Fran^aise de Rome, Rome, pp. 215-218.
2000, "The Hellenistic Economy: Indirect Intervention by the State", in Lo
Cascio and Rathbone, pp. 37-43.
France, J. 1994, "De Burmann a Finley: les douanes dans 1'histoire economique
de 1'empire romain", in Andreau, Briant and Descatet al., pp. 127-153.
1999; "Les revenues douanieres des communautes municipales dans Ie monde
romain", in // capitolo delle entrate nelle finame municipali in Occidente
ed in Oriente, Collection de 1'Ecole Fran^aise de Rome, 256, pp. 95-
113.
2001, Quadragesima Galliarum. L'organisation douaniere des provinces
Alpestres, Gauloises et Germaniques de I'empire Romain, Collection de
1'Ecole Fran^aise de Rome 278, Rome.
2003, "La Ferme des douanes dans les provinces occidentales et 1'empire
romain", in Aubert, pp. 193-214.
Frank, A. G. 1998, ReOrient. Global Economy in the Asian Age, Berkeley and
Los Angeles.
Frank, A. G. and Gills, B. K. (eds. ) 1993, The World System. Five Hundred Years
or Five Thousand? London and New York.
Frank, T. (ed. ) 1933-40, An Economic Survey of Ancient Rome, vols. 1-6,
Baltimore, MD.
Frayn, J. M. 1993, Markets and Fairs in Roman Italy, Oxford.
Frederiksen, M. 1975, "Theory, Evidence and the Ancient Economy", Journal of
Roman Studies 65, pp. 164-171.
Freyberg,H.-U. von 1989, KapitalverkehrundHandelim romischenKaiserreich
('27 v. Chr. -2j5 n. Chr. ), Freiburg.
319
 BIBLIOGRAPHY
Frezouls, E. 1985, "Les ressources de I'euergetisme: Ie cas d'Opramoas de
Rhodiopolis", in Leveau, pp. 249-254.
Friedlander, L. j 92 [-3, Darstellungen aus der Sittengeschichte Roms in der Zeit
van Augustus his 7.um Ausgang der Antonine, vol. 10, ed. Georg Wissowa,
4 vols.. Leipzig.
Frier, Bruce W. 2000a, "Demography", in The Cambridge Ancient History,
vol. XI, 2nd edn, pp. 787-816.
200ob, "More is Worse: Some Observations on the Population of the Roman
Empire", in Scheidel, pp. 139-159.
Fulford, M. 1987, "Economic Interdependence among Urban Communities of
the Roman Mediterranean", World Archaeology 19, I, pp. 58-75.
Fiille, G. 1997, "The Internal Organizationof the Arretine Terra SigillataIndus-
try", Journal of Roman Studies 87, pp. 111-155.
Gabba, E. 1988, Del buon uso delta ricchezw. Saggi di storia economica e sociale
del mondo antico. Milan.
Gabrielsen, V. 1997, The Naval Aristocracy of Rhodes, Aarhus.
2001, "The Rhodian Associations and Economic Activity", in Archibald etal.,
pp. 215-244.
Gagos, T. and van Minnen, P. 1994, Settling a Dispute. Towarda LegalAnthro-
pology of Late Antique Egypt, Ann Arbor, MI.
Galanter,M. 1981,"JusticeinManyRooms:Courts,PrivateOrdering,andIndige-
nous Law", Journal of Legal Pluralism 19, pp. 1-47.
Galsterer, H. 1986, "Roman Law in the Provinces - Some Problems ofTrans-
mission", in Crawford 19860, pp. 13-27.
1996, "The Administration of Justice", in The Cambridge Ancient History,
vol. X, 2nd edn, pp. 397-413.
i997> "Neues zu den romischen Stadtgesetzen", Zeitschrift der Savigny-
Stiftung. Romanistische Abteilung, pp. 392-401.
Gamsey,P. 1970,SocialStatusandLegalPrivilege inthe RomanEmpire,Oxford.
l976> "Urban Property Investment", in Finley, pp. 123-136.
1978, "Rome's African Empire under the Principate" in Gamsey and Whit-
taker, pp. 223-254.
i983a, "Faminein Rome", in Gamsey and Whittaker, pp. 56-65.
i983b, "Grain for Rome", in Garnsey, Hopkins and Whittaker,pp. 118-130.
1988, Famine andFood Supply in the Graeco-RomanWorld, Cambridge.
1991, "The Generosity of Veyne", Journal of Roman Studies 81, pp. 164-
i68.
i994> Foodand Society in Classical Antiquity, Cambridge.
1996, "Prolegomenonto the Studyofthe Landin theLaterRomanEmpire", in
J. H. M. Strubbe, R. A. Tybout and H. S. Versnel (eds. ), Energeia: Studies
on Ancient History and Epigraphy Presented to H. W. Picket, Amsterdam,
pp. 135-153.
i998a. Cities, Peasants and Foodin Classical Antiquity, Cambridge.
l998b, Food and Society in Antiquity, Cambridge.
320
BIBLIOGRAPHY
2000, "The Land", in The Cambridge Ancient History, vol. XI, 2nd edn,
Cambridge, pp. 679-709.
2004, "Roman Citizenship and Roman Law in the Late Empire", in S. Swain
and M. Edwards (eds. ). Approaching Late Antiquity. The Transformation
from Early to Late Empire, Oxford, pp. 133-155.
Garnsey, P., Hopkins, K. and Whittaker, C. R. (eds. ) 1983, Trade in the Ancient
Economy. London.
Gamsey, P. and Humfress, C. 2001, The Evolution of the Late Antique World,
Cambridge.
Garnsey, P. and Saller, R. 1987, The Roman Empire. Economy, Society and
Culture. London.
Garnsey, P. and Whittaker, C. R. 1978 (eds. ). Imperialism in the Ancient World,
Cambridge.
1983 (eds. ), Trade and Faminein Classical Antiquity, Cambridge.
1998, "Trade, Industry and the Urban Economy", in The Cambridge Ancient
History, vol. XIII, 2nd edn, Cambridge.
Geertz, Clifford 1963, Peddlers and Princes. Social Development and Economic
Change in Two IndonesianTowns, Chicago.
i979> "Suq: The Bazaar Economy in Sefrou", in C. Geertz, H. Geertz and L.
Roan, Meaning and Order in Moroccan Society, Cambridge, pp. 123-313.
2000, Local Knowledge. Further Essays in Interpretive Anthropology, New
York.
Gellner, E. i988a, "Introduction", in Baechler, Hall and Mann, pp. 1-5.
i988b. Plough, Sword and Book, Chicago.
Gelzer, M. 1912, Die Nobilitdt der Romischen Republik, Berlin.
Giardina, A. 1986, "Le merci, il tempo, il silenzio: Ricerche su miti e valori
sociali nel mondo Greco e Romano", Studi Storici, 27, 2, pp. 277-302.
(ed. ) 1989, L'uomo romano, Rome and Bari.
l997i L'ltalia Romana. Studi di stone di un'identita incompiuta, Rome.
Giardina, A. and Schiavone, A. (eds. ) 1981, Societa romana e produz. ione schi-
avistica, vols. l-t, Bari and Rome.
Gibbon, E. 1993, The Decline and Fall of the Roman Empire, introduction by
Hugh Trevor-Roper, vols. 1-6, London.
Gibson, A. J. S. andSmout, T. C. 1995,"RegionalPricesandMarketRegions:The
Evolution of the Early Modern Scottish Grain Market", Economic History
Review 48, 2, pp. 258-282.
Glamann, K. 1958, Dutch-Asiatic Trade, 1620-1740, Copenhagen.
i977> "The Changing Patterns of Trade", in The Cambridge Economic History
of Europe, vol. V, Cambridge, chapter4.
Goethe, J. W. van 1907, Maximen und Reflexionen, ecl. Max Hecker, Weimar.
Goitein, S. D. 1967, A Mediterranean Society. The Jewish Communities of the
Arab World as Portrayed in the Documents of the Cairo Geniza, vol. I,
Berkeley and Los Angeles.
Gokhale, B. G. 1979, Surat in the I'jth Century, London.
321
 BIBLIOGRAPHY
Golden, M. and Toohey, P. (eds. ) 1997, Inventing Ancient Culture, London.
Goldschmid, L. 1897, "Les impots et droits de douane en Judee sous les
Remains", Revue des Eludes Juives 34, pp. 192-217.
Goldsmith, R. W. 1987, Pre-Modern FinancialSystems. A Historical Compara-
five Study, Cambridge.
Goldstone, J. 1991, Revolution and Rebellion in the Early Modem World, Berke-
ley, CA.
1998, "The Problem of the 'Early Modern' World", Journal of the Economic
and Social History of the Oriental, pp. 249-284.
2002, "Efflorescences and Economic Growth in World History: Rethinking the
'Rise of the West' and the Industrial Revolution", Journal of World History
13, pp. 323-389.
Gommans, J. 2002, Miighal Warfare. Indian Frontiers and High Roads to Empire,
/500-/700, London and New York.
Gonzales, J. 1986, "The Lex Irnitana: A New Copy of the Flavian Municipal
Law", Journal of Roman Studies 76, pp. 147-243.
Goodman, M. 1983, State and Society in Roman Galilee, AD. 132-212, Totowa,
NJ.
Goody, J. 1990, The Oriental, the Ancient andthe Primitive, Cambridge.
1996, The East in the West, Cambridge.
1998, Food and Love. A Cultural History of East andWest, London.
Goody, J. and Whittaker, D. 2001, "Rural Manufacturing in the Rouergue from
Antiquity to the Present: The Examples of Pottery and Cheese", Compara-
tive Studies in Society and History 43, 2, pp. 225-245.
Gordon, S. 1990, "Burhanpur: Entrepot and Hinterland, , 650-7750", in
Subrahmanyam 19900, pp. 48-65.
Goudineau, C. 1974, "La ceramique dans I'economic romaine", Dossiers
d'Archeologie 6, pp. 103-109.
Granger, C. W. J. and Elliott, C. M. 1967 "A Fresh Look at Wheat Prices and
Markets in the EighteenthCentury", EconomicHistory Review 20, pp. 257-
265.
Greene, K. 1986, The Archaeology of the Roman Economy, London.
i994> "Technology and Innovation in Context: The Roman Background to
Medieval and Later Developments", Journal of Roman Archaeology 7,
PP. 22-33.
2000, "Technological Innovation and Economic Progress in the Ancient World:
M. I. Finley Re-considered", Social and Economic History Review 53, I,
pp. 29-59.
Greif, A. 1993, "Contract Enforceability and Economic Institutions in Early
Trade: The Mahgribi Traders' Coalition", American Economic Review 83,
pp. 525-548.
Gren, E. 1941, Kleinasien und der Ostbalkan in der Wirtschaftlichen Entwick-
lung der romischen Kaiserzeit, Uppsala Universitets Arsskrift 194 r: 9,
Uppsala.
322
BIBLIOGRAPHY
Grenier. J. -Y. 1997, "Economie du surplus, economie du circuit", in Andreau,
Briant and Descat et at., pp. 385-404.
Grover. B. R. 1966, "An Integrated Pattern of Commercial Life in the Rural
Society of North India during the 17th and 18th Centuries", Proceedings
of the 3-/th Session of the Indian Historical Records Commision 38, New
Delhi.
Grlinewald, T. 1999, Rduher, Rebellen, Rivalen, Rdcher. Studien zu Latrones im
romischen Reich, Stuttgart.
Guarino, A. 1972, "Societas consensu contracta". Atti delta Accademia di Scienz.e
Morali e Politiche delta Societa Nationale di Sciem. e. Lettere ed Arti di
Napoli. 83. 261-359.
Gungwu, W. 1990 "Merchants Without Empire: The Hokkien Sojourning Com-
munities", in Tracy, pp. 400-421.
Habib, I. 1963, The Agrarian System of'Mughal India, Bombay.
1969, "Potentialities of Capitalistic Development in the Economy of Mughal
India , Journalof Economic History 29, pp. 32-78.
1972, "The System of Bills of Exchange (Hundis) in the Mughal Empire",
Proceedings of the Indian History Congress, (3j rd session, Muwffarpur),
pp. 290-303.
1985, "Classifying Pre-Colonial India", in Byres and Mukhia, pp. 44-53.
1990. "Merchant Communities in Precolonial India",in Tracy, pp. 371-399.
1999, The Agrarian System ofMughal India, 2nd rev. edn, Oxford and Delhi.
Habib, I. and Raychaudhuri, T. (eds. ) 1982, The Cambridge Economic History
of India, vol. I: c. 1200-c. 1750, Cambridge.
Hackens, R and Miro, M. 1990 (eds. ), Le commerce maritime romain en
Mediterranee occidentale, PACT 27, Strasbourg.
Hack), K. 1997, "Der Zivilprozess des friihen Prinzipats in den Provinzen",
Zeitschriftder Savigny-Stiftung, RomanistischeAhteilung 114. pp. 141-159.
Haider, S. N. 1997, The Monetary System of the Mughal Empire, D. Phil. thesis,
University of Oxford.
Haldon, J. 1993, The State and the Tributary Mode of Production, London.
Halfmann, H. 1979, Die Senatoren aus dem ostlichen Teil des Imperium
Romaniimhis zum Ende des 2. Jahrhundertsn. Chr., Gottingen.
Hall, J. 1985, Powers and Liberties, Oxford.
Hammond, M. 1957, "Composition of the Roman Senate AD 68-235", Journal
of Roman Studies 47, pp. 74-81.
Harl, K. W. 1996, Coinage in the Roman Economy, 300 B. C. -A. D. 700, London.
Harries, Jill 1999, Lawand Empire in Late Antiquity, Cambridge.
Harris, B. 1987, "Merchants and Markets of Grain in South Asia", in Shanin, pp.
205-220.
Harris, W. 1980, "Roman Terracotta Lamps: The Organisationof an Industry",
Journal of Roman Studies 70, pp. 125-145.
1993. "Between Archaic and Modem: Problems in Roman Economic His-
tory", in (ed. ) The Inscribed Economy: Production and Distribution in the
323
 BIBLIOGRAPHY
Roman Empire in the Light of Instnimentum Domesticum, Journal of Roman
Archaeology Supplementary Series 6, Ann Arbor, MI, pp. I [-29.
i999i"Demography, Geography and the Sources of Roman Slaves", Joiirnal
of Roman Studies 89, pp. 62-75.
2000. "Trade and Transport", in The Cambridge Ancient History, vol. XI, 2nd
edn, Cambridge, pp. 710-740.
2003, "Roman Governments and Commerce, 300 B. C. -A. D. 300", in
Zaccagnini, pp. 275-306.
20o6, "A Revisionist View of Roman Money", Journal of Roman Studies 96,
pp. 1-24.
Harris-White, B. (ed. ) 19993, Agricultural Markets From Theory to Practice.
Field Experience in Developing Countries, London.
i999b, "Powerin PeasantMarkets", in Harris-White l999a, pp. 261-286.
Hartwell, R. 1962, "A Revolution in the ChineseIron and Coal IndustriesDuring
the Northern Sung, 960-1126 A. D", Journal of Asian Studies 21, pp. 153-
162.
1966, "Markets, Technology and the Structure of Enterprise in the Develop-
ment of the Eleventh-century ChineseIron and Steel Industries",Journalof
Economic History 26, pp. 29-58.
Hasan, F. 2004, State and Locality in Mughal India. Power Relations in Western
India, c. 1572-1730, Cambridge.
Hassall, M. 2000, "The Army", in The Cambridge Ancient History, vol. XI, 2nd
edn, pp. 320-343.
Haswell, M. 1967, Economics of Development in Village India, London.
Hatzfeld, J. 1919, Les trafiquants italiens dans {'Orient hellenique, Paris.
Heberdey, R. 1897, Opramoas. Inschriften vom Heroon w Rhodiopolis,
Vienna.
Heberdey, R. and Kalinka. E. 1896, Bericht iiber wei Reisen im sudwestlichen
Kleinasien. Vienna.
Hegel, G. W. 1996, Vorlesungen iiber die Philosophie der Welt^eschichte, ed.
K. H. llting, K. Brehmer and H. N. Seelmann, Hamburg.
Herz, P. 1988, Stiidienwr romischen Wirtschaftsgesetzgebung. Die Lebensmit-
telversorgung. Historic! Einzelschriften 55, Stuttgart.
Hiderato, I. 1999-2000, "The Transfer of Money in Roman Egypt", Kodai 10,
pp. 83-104.
Hingley, R. 2005, Globalizing Roman Culture. Unity, Diversity and Empire,
London and New York.
Hinton, W. 1966, Fanshen. A Documentary of Revolution in a Chinese Village,
New York and London.
Hintze, A. 1997, The Mughal Empire and its Decline. An Interpretation of the
Sources of Social Power, Aldershot.
Hitchen, R. B. 2005, '"The Advantages of Wealth and Luxury': The Case
for Economic Growth in the Roman Empire", in Manning and Morris,
pp. 207-222.
324
BIBLIOGRAPHY
Hitchner, R. B. and Mattingly, D. 1995, "Roman Africa: An Archaeological
Review", Journal of Roman Studies 85. pp. 165-213.
Hobenreich, E. 1997, Annona. Juristische Aspekte der Stadtmmischen Lehens-
mittelversorgung in Prinzipat, Graz.
Hobsbawm, E. 1987, The Age of Empire, 1875-1914, London.
Hodgson, M. 1974, The 'Venture of Islam. Conscience and Histon' in a World
Civilization, vols. 1-3, Chicago.
i993i Rethinking World History, ed. Edmund Burke III, Cambridge.
Hont, I. 2005, Jealousy of Trade. International Competition and the Nation-state
in Historical Perspective, Cambridge, MA.
Hopkins, K. 19780, Conquerors and Slaves, Cambridge.
i978b, "Economic Growth and Towns in Classical Antiquity", in P. Abrams
and E. A. Wrigley (eds. ), Towns in Societies. Essays in Economic History
and Historical Sociology, Cambridge, pp. 35-77.
1980, "Taxes and Trade in the Roman Empire 200 B. C. -A. D. 400", Journal
of Roman Studies 70, pp. 101-125.
i983a. Deathand Renewal, Cambridge.
i983b, "Models, Ships and Staples", in Garnsey and Whittaker, pp. 84-109.
19830, "Introduction", in Garnsey, Hopkins and Whittaker, pp. ix-xxv.
1995-6. "Rome, Taxes, Rents and Trade", Kodai. Journal of Ancient Histoi-y
6/7, pp. 41-75.
2000, "Rents, Taxes, Trade and the City of Rome", in Lo Cascio 20ooa,
pp. 253-268.
Hopwood, K. 1989, "Bandits, Elites and Rural Order", in Wallace-Hadrill,
pp. 171-189.
Harden, P. and Purcell, N. 2000, The Corrupting Sea. A Study of Mediterranean
History, Oxford.
Howgego, C. 1992, "The Supply and Use of Money in the Roman World", Journal
of Roman Studies 82, pp. 1-31.
l994> "Coin Circulation and the Integration of the Roman Economy", Journal
of Roman Archaeology 7, pp. 5-21.
i995> Ancient History from Coins, London.
Huang, R. 1974, Taxation and Governmental Finance in Sixteenth-Century Ming
China, Cambridge.
Hultsch, F. 1882, Griechische und romische Metrologie, 2nd rev. edn, Berlin.
Huntington, S. P. 1996, The Clash of Civilizations and the Remaking of World
Order. New York.
Hurst, H. 1993, "Cartagine, la nuova Alessandria", in A. Schiavone (ed. ), Storia
cli Roma, vol. 3, 2, Turin, pp. 327-337.
Husain, M. M. H. (ed. ) 1913, "The Mirza-Namah (The Book of the Perfect
Gentleman) of Mirza Kamran with an English Translation", Journal of the
Asiatic Society of Bengal 9, pp. 1-13.
Inalcik, H. 1969, "Capital Formation in the Ottoman Empire", Journal of Eco-
nomic History 29, pp. 97-140.
325
 BIBLIOGRAPHY
Inalcik, H. and Quataert, D. (eds. ) 1994, An Economic and Social Histor/ of the
Ottoman Empire, vol. 2: i6oo-i()i4. Cambridge.
Inden, R. 1990, Imagining India, Oxford.
Irvine, W. 1962, The Army of the Indian Moghuls. Its Organization and Admin-
istrat'wn. New Delhi.
Islamoglu-tnan, H. (ed. ) 1987, The Ottoman Empire and the 'World-economy,
Cambridge.
Israel, J. I. 1989, Dutch Primacy in World Trade 1585-1740, Oxford.
Jacobsen, G. 1995, Primitiver Austausch oder Freier Markt? Untersuchun-
gen wm Handel in den gallisch-germanischen Provim. en wdhrend der
romischen Kaiserzeit. St Katharinen.
Jacques, F. and Scheid, J. 1990, Rome et I 'integration de I 'empire (44 av. J. . -260
ap. J. C. ). Paris.
Jain, L. C. 1929, Indigenous Banking In India. London.
Jakab.E. 1997, PraedicereundCavere heim Marktkauf. Sachmangelim griech-
ischen und romischen Recht. Munich.
2003, "Pixierte Preise fur die Vermarktung des Weins im Heroninos-Archiv?",
Miinstersche Beitrdge wr Antiken Handelsgeschichte22, l, pp. 1-26.
2006, "Vertragsformulare im Imperium Romanum", Zeitschrift der Savigny-
Stiftungfiir Rechtsgeschichte, Romanistische Abteilung 123, pp. 71-101.
Jeannin,J. P. 1982, "Les interdependanceseconomiques dans Ie champ d'action
europeen des Hollandais (XVIe-XVIIesiecle), in Aymard, pp. 147-170.
Johnston, A. C. 1936. Roman Egypt to the Reign of Diocletian. Baltimore, MD.
Johnston, D. 1999, Roman Law in Context, Cambridge.
Jones, A. H. M. 1940. The Greek City from Alexander to Justinian, Oxford.
1964, The Later Roman Empire 284-602. A Social, Economic and Adminis-
trative Survey, vols. 1-2. Norman. OK.
1974; The Roman Economy, ed. P. A. Brunt, Oxford.
Jones, D. 2006, The Bankers of Puteoli. Finance, Trade and Industry in the
Roman World. Stroud.
Jones, E. L. 1981,The European Miracle, Cambridge.
1988, Growth Recurring. Economic Change in World History, Oxford.
Jongman, W. 1988, The Economy and Society of Pompeii, Amsterdam.
2000a, "Hunger and Power: Theories, Models and Methods in Roman Eco-
nomic History", in H. Bongenaar(ed. ), Interdependency of Institutions and
Private Entrepreneurs, MOS Studies 2, Leiden,pp. 259-284.
2000b,"WoolandtheTextileIndustryofRomanItaly:A WorkingHypothesis",
in Lo Cascio 20ooa, pp. 187-198.
2006, "The Rise and Fall of the Roman Economy: Population, Rents and
Entitlement", in Bang, Ikeguchiand Ziche, pp. 237-254.
Kajanto, L. 1965,The LatinCognomina,SocietasScientiarumFennica,Helsing-
fors.
Kaser, M. 1996, Das romische Zivilprozessrecht, 2nd edn, rev. by K. Hackl,
Munich.
326
BIBLIOGRAPHY
Katzoff, R. 1972, "Precedents in the Courts of Roman Egypt", Zeitschnft der
Savigny-Stiftung, Romanistische Abteilung 89, pp. 256-292.
Kautsky, J. H. 1982, The Politics of Aristocratic Empires, Chapel Hill, NC.
Keay, S. J. 1984, Late Roman Amphorae in the Western Mediterranean. A Typol-
iigy and Economic Study: The Catalan Evidence, BAR International Series
n6, Oxford.
(ec). ) 1998, The Archaeology of Early Roman Baetica, Journal of Roman
Archaeology, Supplementary Series 29, Ann Arbor, MI.
Kehoe, D. 1993, "Economic Rationalism in Roman Agriculture", Journal, of
Roman Archaeology 6. pp. 476-484.
J997' Investment, Profit and Tenancy. The Jurists and the Roman Economy,
Ann Arbor. Ml.
Kennedy, P. 1988,The RiseandFall ofthe Great Powers. Economic Changeand
Military Conflict from ,500 to 2000. London.
Kenrick, Philip 1993. "ItalianTerra Sigillata: A SophisticatedRoman Industry",
OxfordJournalof Archaeology 12, 2, pp. 35-42.
2000, Corpus Vasorum Arretinorum, 2nd edn., Bonn.
Kessler, D. andTemin, P. 2005, "Money andPricesin the Early Roman Empire",
MIT Department of Economics Working Paper No. 05-11, Cambridge,
MA.
Keyder, C. 1997, "TheOttoman Empire", in K. BarkeyandM. von Hagen(eds. ),
After Empire, Boulder, CO.
Kirschenbaum, A. \c)%l, Sons, Slaves and Freedmen in Roman Commerce, Wash-
ineton. DC.
Klein, P. W. 1965, De Trippen in de /ye eeuw. En studie over het ondernemers-
gedrag op de Hollandse stapelmarkt, Assen.
Klingenberg, G. 1977, Commissum. Der Vervall nichtdeklarierter Sachen im
rdmischen Zollrecht, Graz.
Kloft, H. 1996, "Uberlegungen zum Luxus in der frtihen romischen Kaiserzeit",
in J. H. M. Strubbe,R. A. Tybout andH. S. Versnel (eds. ), Energeia: Studies
on Ancient History and Epigraphy Presented to H. W. Picket, Amsterdam,
pp. 113-134.
Kloppenborg, J. S. and Wilson, S. G. (eds. ) 1996, Voluntary Associations in the
Graeco-Roman World, London and NewYork.
Kneissl, P. 1998, "Die Berufsvereine im romischen Gallien", in P. Kneissl and V.
Losemann(eds. ), Im.periumRomanum. FestschriftfiirKarlChrist, Stuttgart,
pp. 431-449.
Kobes, J. 1999, "Fremdes Getreide: Beobachtungenzum Problem der Getreide-
versorgung in der kaiserzeitlichen ProvinzAsia", Laverna 10, pp. 81-98.
Kokkinia, C. 2000, Die Opramoas-Inschrift von Rhodiapolis, Bonn.
Krause, J. -U. 1994, Witwen und Waisen im rdmischen Reich, II: Wirtschaftliche
und gesellschaftliche Stellung van Witwen, Stuttgart.
l995> Wihven und Waisen im rdmischen Reich, III: Rechtliche und soT. iale
Stellung van Waisen, Stuttgart.
327
 BIBLIOGRAPHY
Kruit, N. 1992, "Local Customs in the Formulas of Sales of Wine for Future
Delivery", Zeitschrift fur Papyrologie und Epigraphik 94, pp. 167-184.
Kruit, N. and Worp, A. 1999, "Metrological Notes on Measures and Containers
of Liquids in Graeco-Roman and Byzantine Egypt", Archiv fur Papyrus-
forschung 45, pp. 96-127.
Kudlien, F. 1994, "Die Rolle der Konkurrenz im antiken Gesellschaftsleben",
Miinstersche Beitrdge ziir Antiken Handelsgeschichte 13, I, pp. 1-39.
1997, "Der antike Makler - ein verleugneter Beruf, Miinstersche Beitrcige zur
Antiken Handelsgeschichte 16, I, pp. 67-84.
Kula, W. 1976, An Economic Theory of the Feudal System, London.
Laet, S. de 1949, Portorium. Etude sur I'organisation douaniere chez les
Romains, Bruges.
Landes, D. 1998, The Wealth and Poverty of Nations. Why Some Are So Rich
and Some So Poor. New York.
Lane, F. C. 1966, Venice and History. The Collected Papers of Frederic C. Lane,
Baltimore, MD.
Larsen M. T. 1967, Old Assyrian Caravan Procedures, Istanbul.
1976, The Old Assyrian City-State and its Colonies, Mesopotamia 4, Copen-
hagen.
Laurence, R. 1998, "Land Transport in Roman Italy: Costs, Practice and the
Economy", in Parkins and Smith, pp. 129-148.
i999i The Roads of Roman Italy, London.
Lendon, J. E. 1997, Empire of Honour, Oxford.
Lepelley, C. (ed. ) 1998, Rome et I'integrationde I'empire (44 av. J. C. - 260 ap.
J. C), tome 2: Approches regionales du Haut-Empire romain, Paris.
Lesger, Cle 2006, The Rise of the Amsterdam Market and Information Exchange:
Merchants, Commercial Expansion and Change in the Spatial Economy of
the Low Countries, c. ,550-7630, Ashgate.
Leur, J. C. van 1955, Indonesian Trade and Society. Essays in Asian Social and
Economic History, The Hague.
Leveau, P. (ed. ) 1985, L'Origine des richesses depensees dans la ville antique,
Aix-en-Provence.
Levi-Strauss, C. 1962, La Pensee sauvage, Paris.
Levy-Bruhl, L. 1910, Les Fonctions mentale dans les societes inferieures, Paris.
1985, How Natives Think, with an introduction by C. Scott Littleton, London
(first published 1926).
Lewis, N. (ed. ) 1989, The Documents from the Bar Kokhba Period in the Cave
of Letters. Greek Papyri, Judean Desert Studies, Jerusalem.
Liebenam,W. 1900, Stddtevenvaltung im rdmischen Kaiserreiche,Leipzig.
Ligt, L. de 1993, Fairs and Markets in the Roman Empire. Economic and Social
Aspects of Periodic Trade in a Pre-Industrial Society, Amsterdam.
2000, "Governmental Attitudes towards Markets and Collegia", in Lo Cascio
20ooa, pp. 237-252.
2002, "Tax Transfers in the Roman Empire", in de Blois and Rich, pp. 84-102.
328
BIBLIOGRAPHY
Lintott, A. 1968, Violence in Republican Rome, Oxford.
I993> Imperium Romanum. Politics and Administration, London.
Liou, B. and Rodnguez-Almeida, E. 2000, "Les Inscriptions des amphores du
Pecio Gondolfo", Melanges de I'Ecole Francaise de Rome. Antiquite 112,
I, pp. 7-25.
Liou, B. and Tchemia. A. 1994, "L'interpretation des inscriptions sur les
amphores Dressel 20", in Epigrafia della produzione e delta distribuwme,
pp. 133-156.
Littleton, C. S. 1985, "Introduction", in Levy-Bruhl, pp. iii-lviii.
Livi-Bacci, M. 1992,A Concise History of World Population, trans. Carl Ipsen,
2nd edn. Oxford.
Lo Cascio, E. 1991, "Forme dell'economia imperiale", in A. Schiavone (ed. ),
Storia di Roma, vol. 2, 2, pp. 313-365.
1998. "Considerazioni su circolazione monetaria, prezzi e fiscalita nel IV
secolo", in G. Crifo and S. Giglio (eds. ), Atti dell'Accademia Romanistica
Costantiniana. XII Convegno Internazionale, Naples, pp. 121-136.
(ed. ) 2000a, Mercati Pennanentie Mercati PeriodicinelMondo Romano. Atti
degli Incontri capresi di storia dell'economia antica, Bari.
2000b, // Princeps e il suo impero. Studi di storia amministrativa efinan7. ia. na
romana, Bari.
2000C, "The Roman Principate: The Impact of the Organization of the Empire
on Production", in Lo Cascio and Rathbone, pp. 77-85.
2003a, "Appaltatori delle imposte e amministrazione finanziaria imperiale",
in Aubert pp. 249-268.
(ed. ) 2003b, Credito e moneta nel mondo romano, Bari.
2003C, "Mercato libero e 'commercio amministrato' in eta tardoantica", in
Zaccagnini, pp. 307-325.
2006, "The Role ofthe State in the Roman Economy: MakingUse ofthe New
Institutional Economics", in Bang, Ikeguchi and Ziche, pp. 215-234.
Lo Cascio, E. and Rathbone, D. (eds. ) 2000, Production and Public Powers in
Classical Antiquity, The Cambridge Philological Society Supplement 26,
Cambridge.
Lopez, R. S. and Raymond, I. W. 1955, Medieval Trade in the Mediterranean
World. Illustrative Documents Translated with Introductions and Notes,
London.
Love, J. R. 1991, Antiquity and Capitalism. Max Weber and the Sociological
Foundations of Roman Civiliwtion, London.
Lund, J. 1999, "Trade Patterns in the Levant from c. 100 BC to AD 200
as Reflected by the Distribution of Ceramic Fine Wares in Cyprus",
Munstersche Beitrdge zur Antiken Handelsgeschichte 18, I, pp. 1-22.
MacMullen, R. 1974, Roman Social Relations. 50 B. C. -A. D. 284, New Haven,
CT.
I99°> Changes in the Roman Empire. Essays in the Ordinary, New Haven, CT.
Maddison, A. 2001, The World Economy. A Millennial Perspective, Paris.
329
 ^rr
BIBLIOGRAPHY
Maine. H. S. i86i. Ancf'en/ Law. Gloucester.
1871, Village Communities in the East and West, London.
Malinowski, B. 1924, Argonauts of the Western Pacific. London.
1948, Magic, Science and Religion and Other Essays, New York.
Malmendier, U. 2002, Societas Puhlicanorum, Cologne.
Manacorda,D. 1989,"Leanforedell'Italiarepubblicana",in Amplwres romaines
el histoire economique. Dix ans de recherche. Collection de 1 Ecole
Fran^aisede Rome [ 14, Rome pp. 443-467.
Mann, M. 1986, The Sources of Social Power. A History of Power from the
Beginning to A. D. 1760, Cambridge.
Manning, J. G. and Morris, I. (eds. ) 2005, The Ancient Economy. Evidence and
Models, Stanford, CA.
Markovits, C. 2000, The Global World of Indian Merchants, , 750-79^7. Traders
of'Sindfrom Bukhara to Panama, Cambridge.
Marsh, G. 1981, "London's Samian Supply and its Relationship to the Devel-
opment of the Gallic Samian Industry", in Anderson and Anderson, pp.
173-238.
Marshall, A. 1890, Principles of Economics, London.
Marshall,A. J. 1980,"TheSurvivalandDevelopmentofInternationalJurisdiction
in the Greek World", Aufstieg imd Niedergang der Rdmischen Well 2, 13,
pp. 626-661.
Marshall, P. J. 1976, "British Merchants in Eighteenth-Century Bengal", Bengal
Past and Present 86, pp. 151-163.
1993, "Retrospect on J. C. Van Leur's Essay on the Eighteenth Century as a
Category in Asian History", Itinerario 17, I, pp. 45-58.
Marx, K. and Engels, F. 1960, The First Indian. War of Independence, London.
Masschale, J. 1993, "Transport Costs in Medieval England", Economic History
Review46, pp. 266-279.
Masters, B. 1988, The Origins of Western Economic Dominance in the Middle
East. Mercantilism and the Islamic Economy. 1600-1750, New York.
Matthews, J. P. 1984, "The Tax Law of Palmyra: Evidence for Economic History
in a City of the Roman East", Journal of Roman Studies 74, pp. 157-180.
Mattingly, D. J. i988a,"0il for Export? A Comparison ofLibyan, Spanish and
Tunisian Olive Oil Production in the Roman Empire", Journal of Roman
Archaeology I, pp. 33-56.
i988b, "The Olive Boom: Oil Surpluses, Wealth and Power in Roman Tripoli-
tania", Libyan Studies 19, pp. 21-41.
1996,"FirstFruit?TheOlivein theRomanWorld",in G. ShipleyandJ.Salmon
(eds. ), Human Landscapes in the Ancient World. Environment and Culture,
London, pp. 213-253.
i997a; "Beyond Belief? Drawing a Line Beneath the Consumer City", in
Parkins i997b, pp. 210-218.
(ed. ) i997b. Dialogues in Roman Imperialism,JournalofRomanArchaeology
Supplementary Series 26, Ann Arbor, MI.
330
BIBLIOGRAPHY
i997c- "Africa, a Landscape of Opportunity", in Mattingly l99?b, pp. 117-
139.
Mattingly, D. J. and Salmon, J. (eds. ) 2001, Economies Beyond Agriculture in
the Classical World, London.
Mattingly, D. J.. Stone, D., Stirling, L. and Ben Lazreg, N. 2001, "Leptiminus
(Tunisia): a 'Producer' City", in Mattingly and Salmon, pp. 66-98.
Mauro, F. 1990, "Merchant Communities, 1350-1750", in Tracy, pp. 255-286.
McKenclrick,N., Brewer,J. andPlumb,J.H. (eds. ) 1982,The Birthofa Consumer
Society.The CommercializationofEighteenth-CentiiryEngland,Blooming-
ton, IN.
McNeill, W. H. 1982, The Pursuit of Power, Chicago.
1986, Mythistory and Other Essays, Chicago.
McPherson, K. 1993, The Indian Ocean. A History of People and the Sea, Delhi.
Meiggs, R. 1973, Roman Ostia, 2nd edn., Oxford.
Meijer, F. and van Nijf, 0. 1992, Trade. Transport, and Society in the Ancient
World. A Sourcebook. London.
Meilink-Roelofsz, M. A. P. 1962, Asian Trade and European Influence in the
Indonesian Archipelago between ,500 and about 1630, The Hague.
1980, "The Structure of Trade in Asia in the Sixteenth and Seventeenth Cen-
turies", in Mare Luso-Indiciim, IV, Paris, pp. 1-43.
Merola, G. D. 2001, Autoiwmia locale governo imperiale, Bari.
Meyer, Ecluard 1910, Einleitung. Elemente der Anthropologie. Geschichte des
Altertums. vol. i. I. Berlin.
1924, "Die Wirtschaftliche Entwicklung des Altertums", in E. Meyer, Kleine
Schriften, vol. I, 2nd edn, Halle, pp. 79-168.
Meyer Elizabeth, A. 2004, Legitimacy and Law in the Roman World, Tabulae in
Roman Belief and Practice, Cambridge.
Mickwitz, G. 1936, Die Kartellfunktionen der Zimfte und ihre Bedeutung bei der
Entstehung des Zunftwesens, Helsinki.
1937) "Economic Rationalism in Greco-Roman Agriculture", English Histor-
ical Review 52, pp. 577-589.
Middleton, P. 1983, "The Roman Army and Long Distance Trade", in Garnsey
and Whittaker, pp. 75-83.
Millar, F. 1964, A Study ofCassius Dio, Oxford.
1977; The Emperor in the Roman World, London.
1981, "The World of the Golden Ass", Journal of Roman Studies 71,
pp. 63-75.
1993' The Roman Near East, j/ BC-AD 337, Cambridge, MA.
1998, "Caravan Cities: The Roman Near East and Long-Distance Trade by
Land, " in M. Austin, J. Harries and C. Smith (eds. ), Modus Operandi.
Essays in Honour of Geoffrey Rickman. Bulletin of the Institute of Classical
Studies 71, Supplement, London, pp. 119-137.
i999> "The Greek East and Roman Law: The Dossier of M. Cn. Licinius
Rufinus", Journal of Roman Studies, 89, pp. 90-108.
331
 BIBLIOGRAPHY
Miller, J. I. 1969, The Spice Trade of the Roman Empire, Oxford.
Millett, M. 1990, The Romaniwtion of Britain. An Essay in Archaeological
Interpretation, Cambridge.
Millett, P. 2001, "Productive to Some Purpose? The Problem of Ancient Eco-
nomic Growth", in Mattingly and Salmon, pp. 17-48.
Minnen, P. van 1986, "The Volume of the Oxyrhynchite Textile Trade ,
Miinstersche Beitrdgewr Antiken Handelsgeschichte 5. 2, pp. 88-95.
1987, "UrbanCraftsmen in Roman Egypt", Miinstersche Beitrdge ziirAntiken
Handelsgeschichte 6, I, pp. 31-88.
Mirat-i-Ahmadi. A Persian History ofGujarat, by Ali Muhammad Khan, trans.
M. F. Lokhandwala, Baroda 1965.
Mitchell, S. 1993, Anatolia: Lcind, Men and Gods in Asia Minor, vols. 1-2,
Oxford.
Mitteis, L. 1891, Reichsret und Volksrecht in den ostlichen Provinzen des
rmnischen Kaiserreichs. Mit Beitrdgen wr Kenntnis des griechischen
Rechts und der spatrdmischen Rechtsentwicklung, Leipzig.
Mitteis, L. and Wilcken, U. 1912, Grunwge und Chrestomathie der
Papyruskunde, vols. I-II, Leipzigand Berlin.
Modrzejewski, J. 1970, "La regle de droit dans 1'Egypte romaine", Proceedings
of the XIIthInternational Congress ofPapyrology, Toronto, pp. 317-377-
1990. Droit imperial et traditions locales dans I'Egypte romaine, Aldershot.
Mommsen, T. 1843, De collegiis et sodaliciis Romanorum, Kiel.
[904, Romische Geschichte, vol. V, Berlin.
1992, Romische Kaisergeschichte nach den Vorlesungen - Mitschriften van
Sebastianund Paul Hensel 1882-86, ed. Barbara and Alexander Demandt,
Munich.
Montesquieu, C. 1955, De I'esprit des lois. Texte etahli et presente par Jean
Brethe de la Gressaye. Paris.
Moosvi, S. 1986-7,"TheGrossNationalProductoftheMughalEmpire,c. 1600",
Indian Historical Review 13, 1-2, pp. 75-87.
1987, The Economy of the Mughal Empire, c. 1595, Delhi.
2000, "The IndianEconomicExperience",in Panikkar,Byres and Patnaik,pp.
328-358.
Morel, Jean Paul 1985, "La Manufacture. Moyen d'enrichissement dans 1'Italie
romaine", in Leveau, pp. 87-112.
Morley, Neville 1996, Metropolis and Hinterland. The City of Rome and the
Italian Economy 200 B. C. -A. D. 200, Cambridge.
2000, "Markets, Marketing and the Roman Elite", in Lo Cascio 2000a, pp.
21 I-223.
2007,TradeinClassicalAntiquity. KeyThemesinAncientHistory,Cambridge.
Morris, I. 2002, "Hard Surfaces", in Cartledge, Cohen and Foxhall, pp. 8-43.
Mouritsen, H. 1988, Elections, Magistrates and Municipal Elite. Studies in Pom-
peian Epigraphy, Analecta Romana Instituti Daniel, Suppl. 15, Rome.
332
BIBLIOGRAPHY
Nafissi, M. 2005, Ancient Athens and Modern Ideology. Value, Theory and Evi-
dence in Historical Sciences. Max Weher, Karl Polanyi and Moses Finley,
London.
Needham,Joseph 1954, Science and Civilization in China,vol. I, Cambridge.
Neesen, L. 1980, Untersuchungen w den direkten Staatsabgahen der romischen
Kaiserzeit (27 v. Chr-284 n. Chr. ), Bonn.
Neeve, P. W. de 1985, "The Price of Agricultural Land in Roman Italy and the
Problem of Economic Rationalism", Opus 4, pp. 77-109.
Nicolet, C. 1980, "Economie, societe et institutions au lie siecle av. J.-C. : De
la lex Claudia a 1'ager exceptus", Annales, Histoire, Sciences Sociales, pp.
871-894.
1988, Rendre a Cesar. Economic et societe dans la Rome antique, Paris.
1991 a. Geography, Space and Politics in the Early Roman Empire, Ann Arbor,
Ml.
i99lb, "Le 'Monumentum Ephesenum' et les dimes d'Asie", Bulletin de Cor-
respondance Hellenique 115, pp. 465-480.
i993i "Le 'Monumentum Ephesenum' et la delimitation duportorium d'Asie ,
Melanges de I'Ecole Fran^aise de Rome. Antiquite 105, pp. 929-959.
1996, Financial Documents and Geographical Knowledge in the Roman
World. The Sixteenth J. L. Mvres Memorial Lecture. Oxford.
Nijf, 0. M. van 1997,The CivicWorldofProfessionalAssociationsin the Roman
East. Amsterdam.
Nippel, W. 1990, Griechen, Barbaren und "Wilde". Alte Geschichte und
Sozialanthropologie, Frankfurt.
1995, Public Order in Ancient Rome, Cambridge.
Nolle, J. 1999, "Marktrechte ausserhalbder Stadt: Lokale Autonomie zwischen
Statthalter und Zentralort", in Eck, pp. 93-114.
Non-, D. 1998,"RomischesZivilprozessrechtnachMax Kaser: Prozessrechtund
Prozesspraxis in der Provins Arabia", Zeitschrift der Savigny-Stiftung fur
Rechtsgeschichte. Romanistische Abteilung 115, pp. 80-98.
North, D. C. 1977, "Markets and Other Allocation Systems in History: The
Challenge of Karl Polanyi", Journal of European Economic History 6, 3,
pp. 703-716.
1990, Institutions, Institutional Change and Economic Performance, Cam-
bridge.
2005, Understanding the Process of Economic Change, Princeton, NJ and
Oxford.
North, D. C. and Thomas, R. P. 1973, The Rise of the Western World. A New
Economic History, Cambridge.
Noy, D. 2000, Foreigners at Rome. Citizens and Strangers, Swansea.
N0rlund, P. 1920, Det romerske slavesamfund under afvikling, Copenhagen.
O'Brien, P. K. 1998, "Inseparable Connections: Trade, Economy, Fiscal State,
and the Expansion of Empire, 1688-1815", in P. J. Marschall (ed. ), Oxford
333
 BIBLIOGRAPHY
History of the British Empire, vol. 2: The EighteenthCentury, Oxford, pp.
53-77.
Obeysekere, G. 1992, The Apotheosis of Captain Cook. European Mythmaking
in the Pacific, Princeton, NJ.
Oliver, J. H. 1953, "The Ruling Power: A Study of the Roman Empire in the
Second Century After Christ Through the Roman Oration of Aelius Aris-
teides". Transactions of the American Philosophical Society n. s. 43, 4,
pp. 871-1003.
1989, Greek Constitutions of Early Roman Emperors from Inscriptions and
Papyri, Philadelphia,PA.
Osborne, R. 1991, "Pride and Prejudice, Sense and Subsistence: Exchange and
Society in the Greek City". in Rich and Wallace-Hadrill, pp. 119-145.
[996> "Pots, Trade and the Archaic Greek Economy", Antiquity 70, pp. 31-
44.
Paludan, H. 1995, Familia og Familie. To europceiske kulturelementers mipde i
hftjmiddelalderens Danmark, Aarhus.
1996, "Pots, Trade and the Archaic Greek Economy", Antiquity 70, pp. 31-44.
Pamuk, 0. 2005, Istanbul. Memories ofthe City, trans. MaureenFreely, London.
Pamuk, S. 2000, "The Ottoman Empire in the Eighteenth Century", Itinerario
24, pp. 104-116.
Panella, C. 1981, "La clistribuzione e i mercati", in Giardina and Schiavone, pp.
55-8o.
1993; "Merci e scambi nel Mediterraneo tardoantico", in A. Schiavone (ed. ),
Storia di Roma, vol. 3, 2, Turin, pp. 613-697.
Panella, C. and Tchernia, A. 1994, "Produits agricoles transportes en amphores:
L'huile et surtout Ie vin" in L'ltalie d'Au^uste a Diocletien, Collection de
I'Ecole Fran^aise de Rome 198, Rome, pp. 145-165.
Panikkar, K. N., Byres, T. J. and Patnaik, U. (eds. ) 2000, The Making of History.
Essays Presented to Irfan Hahib, New Delhi.
Parker, A. J. 1992, Ancient Shipwrecks of the Mediterranean and the Roman
Provinces, BAR International Series 580, Oxford.
1996, "Sea Transport and Trade in the Ancient Mediterranean", in E, E. Rice
(eds. ). The Sea and History, Gloucester, pp. 97-109.
Parkins, H. i997a, "The 'Consumer-city' Domesticated? The Roman City in
Elite Economic Strategies", in Parkins i997b, pp. 83-111.
(ed. ) i997b, Roman Vrbanism Beyondthe Comsumer City, London.
1998, "Time for Change? Shaping the Future of the Ancient Economy", in
Parkins and Smith, pp. 1-15.
Parkins, H. and Smith, C. 1998 (eds. ). Trade, Traders and the Ancient City,
London.
Parthasarathi, P. 2002, "Review Article: The Great Divergence", Past and Present
176, pp. 275-293.
Paterson, J. 1998, "Trade and Traders in the Roman World: Scale, Structure, and
Organization", in Parkins and Smith, pp. 149-167.
334
BIBLIOGRAPHY
Patterson. J. R. 1987, "Crisis: What Crisis? Rural Change and Urban Development
in Imperial Apennine Italy", Papers of the British School at Rome 55, pp.
115-146.
[ 99[- "Settlement, City and Elite in SamniumanclLycia",in RichandWallace-
Hadrill. pp. 147-168.
1994. "The Collegia and the Transformation of the Towns of Italy in the
Second Century AD", in L'ltalie d'Augitste a Diocletien. Collection de
I'Ecole Frangaise de Rome 198, Rome, pp. 227-238.
2006, Landscapes and Cities. Rural Settlement and Civic Transformation in
Early Imperial Italy, Oxford.
Pavis d'Escurac, H. 1976, La Prefectiire de I'annone, service administratif
imperial d'Augiiste a Constantin, Rome.
1977. "Anstocratie senatoriale et profits commerciaux", Kteina 2, pp. 339-
355.
Peacock, D. P. S. and Williams. D. F. 1986, Amphorae and the Roman Economy,
London.
Pearson, M. N. [976, Merchants and Rulers in Gujarat. The Response to the
Portuguese in the Sixteenth Century, Berkeley, CA.
1985, "Land, Noble, and Ruler in Mughal India", in E. Leach, S. N. Mukherjee
and J. 0. Ward (eds. ), Feiidalism. Comparative Studies, Sydney, pp. 175-
196.
i988a. Before Colonialism. Theories of Asian-European Relations, New
Delhi.
tQSSb,"Brokers in WesternIndianPort Cities:TheirRolein ServicingForeign
Merchants", Modern Asian Studies 22, 3, pp. 445-472.
1991- "Merchants and States", in Tracy 1991, pp. 41-1i6.
1998, Port Cities and Intruders. The Swahili Coast, India, and Portugal in the
Earlv Modern Era. Baltimore. MD and London.
Pearson, R. 1993, "Taking Risks and Containing Competition: Diversification
and Oligopoly in the Fire Insurance Markets in the North of England During
the Early Nineteenth Century", Economic History' Review 46, I, pp. 39-
64.
Pelsaert, F. 1925, Jahungir's India. The Remonstrantie of Francisco Pelsaert,
translated from the Dutch by W. H. Moreland and P. Geyl, Cambridge.
Perkins, D. 1969, Agricultural Development in China, 1368-1968, Chicago.
Perlin, F. 1978, "Of White Whale and Country-men in the Eighteenth-century
Maratha Deccan", Journalof Peasant Studies 5, 2, pp. 172-243.
1981, "The Pre-Colonial Indian State in History and Epistemology", in
H. Claessen and P. Skalnik (eds. ). The Study of the State, The Hague,
pp. 275-302.
1983, "Proto-Industrializationand Pre-Colonial South Asia", Past andPresent
98, pp. 30-95.
[993> The Invisible City. Monetary, Administrative and Popular Infrastruc-
tures in Asia and Europe, lyM -igoo, Aldershot.
335
 BIBLIOGRAPHY
1994. "Changes in Production and Circulation of Money in Seventeenth and
Eighteenth Century India: An Essay ofMonetization Before Colonial Occu-
pation", in Subrahmanyam,pp. 276-306.
Persson, K. G. 1988, Pre-Industrial Economic Growth. Social Organization and
Technological Progress in Europe, Oxford.
i999> Grain Markets in Europe , 500-7900. Integration and Deregulation,
Cambridge.
Pflaum, H. G. 1950, Les Procurateurs equestres sous Ie Haut-Empire romain,
Paris.
1960, Les Carrieres procuratoriennes equestres sous Ie Haut-Empire romain,
vols. 1-3, Paris.
Picket, H. 1983, "Urban Elites and Business in the Greek Part of the Roman
Empire", in Garnsey, Hopkins and Whittaker, pp. 131-144.
1984, "Urban Elites and the Economy in the Greek Cities of the
Roman Empire", Miinstersche Beitrcige zur Antiken Handelsgeschichte I,
pp. 3-36.
[990, "Wirtschaft", in F. Vittinghoff (ed. ), Europdische Wirtschqfts- und
Sozialgeschichte in der romischen Kaiserzeit, Stuttgart, pp. 25-160.
J 994, "The Roman State and the Economy: The Case ofEphesus", in Andreau,
Briant and Descat et al., pp. i15-126.
Pocock, J. G. A. 2003, Barbarism and Religion, vol. 3: The First Decline and
Fall, Cambridge.
Polanyi, K. 1957, "The Economy as Instituted Process", in Polanyi, Arensberg
and Pearson, pp. 243-270.
1963, "Ports of Trade in Early Societies", Journal of Economic History 23, I,
pp. 30-45.
I977> The Livelihood of Man, London.
Polanyi, K., Arensberg, C. and Pearson, H. (eds. ) 1957, Trade and Market in the
Early Empires, New York.
Pollard, N. 2000, Soldiers, Cities, and Civilians in Roman Syria, Ann Arbor,
MI.
Pollock, S. 1998, "India in the Vernacular Millennium: Literary Culture and
Polity", Daedalus 127, 3, pp. 41-75.
Pomeranz, K. 2000, The Great Divergence. China, Europe and the Making of
the Modern World Economy. Princeton. NJ.
Pomey, P. and Tchernia, A. 1977, "Le tonnage maximum des navires de com-
merce romains", Archaeonautica 2, pp. 233-251.
Ponsich, M. 1998, "The Rural Economy of Western Baetica", in Keay, pp. 171-
l82.
Postan, M. M. 1972, The Medieval Economy and Society. An Economic History
of Britain in the Middle-Ages, Birkenhead.
Posthumus, N. W. 1946, Inquiry into the History of Prices in Holland, Leiden.
Prachner, G. 1980, Die Sklaven und Freigelassenen im Arretinischen Sigil-
latengewerbe, Wiesbaden.
336
BIBLIOGRAPHY
Prakash, 0. 1972, "The Dutch East India Company in Bengal: Trade Privileges
and Problems, 1633-1712", Indian Economic and Social History Review 9,
pp. 258-287.
1985,The DutchEast IndiaCompany andthe Economy of Bengal, 1630-1720,
Princeton, NJ.
1991. "Precious Metal Flows, Coinage and Prices in India in the i7th and
the Early 18th Century", in E. van Cauwenberghe (ed. ), Money, Coins and
Commerce. Essays in the Monetary History of Asia and Europe, Louvain,
PP. 55-73.
(ed. ) 1997,EuropeanCommercialExpansionin Early ModemAsia. AnExpan-
ding World, vol. 10, Aldershot.
1998, European Commercial Enterprise in Pre-Colonial India, The New Cam-
bridge History of India 2, 5, Cambridge.
Preisigke, F. 1910, Der Girowesen im griechischen Agypten, Strasbourg.
Pringsheim, F. 1950, The Greek Law of Sale, Weimar.
Ptak, R. and Rothermund, D. (eds. ) 1991, Emporia, Commodities and Entre-
preneurs in Asian Maritime Trade, c. 1400-1750, Stuttgart.
Purcell, N. 2005, "The Ancient Mediterranean: The View from the Customs
House", in W. V. Harris (ed. ), Rethinking the Mediterranean, Oxford, pp.
200-232.
Puqiura, G. 1985, "II regolamento doganale di Cauno e la lex Rhodia in D.
I4. 2. 9", Annali Seminario Giuridico della Universita di Palenno 38, pp.
273-331.
Qaisar, A. J. 1967, "Distribution of the Revenue Resources of the Mughal Empire
among the Nobility", in Proceedings of the Twenty-Seventh Session of the
Indian History Congress, Aligarh, pp. 237-243.
Quass,F. 1993, Die Honoratiorenschichtin denStddtendesgriechischenOstens.
Untersuchungen 7.ur politischen undso7. ialen Entwicklung in hellenistischer
und romischer Zeit, Bonn.
Quereschi, J. H. 1966, The Administration of the Mughal Empire, Karachi.
Rapsaet, G. 1987, "Aspects de 1'organisation du commerce de la ceramique
sigillee dans Ie Nord de la Gaule au lie siecle de notre ere, I: Les donnees
materielles", Miinstersche Beitrdge wr antiken Handelsgeschichte 6, 2,
pp. 1-29.
Rapsaet-Charlier,M.-T. andG. 1988,"Aspectsde 1'organisationdu commerce de
la ceramique sigillee dans Ie Nord de la Gaule au lie siecle de notre ere I: Les
donnees materielles", Munstersche Beitrdge zur antiken Handelsgeschichte
7, 2, pp. 45-69.
Raschke, M. G. 1978, "New Studies in Roman Commerce with the East",
Aufstieg und Niedergang der romischen Welt 2, 9. 2, pp. 604-1378.
Rathbone, D. 19830, "The Weight and Meassurement of Egyptian Grains",
Zeitschrift fiir Papyrologie und Epigraphik 53, pp. 265-275.
l983b, "Italian Wines in Roman Egypt", Opus 1, pp. 81-98.
337
 BIBLIOGRAPHY
1991, Economic Rationalism and Rural Society in Third-Centw-y A. D. Egypt,
Cambridee.
1997' "Prices and Price Formation in Roman Egypt", in Andreau, Briant and
Descat et at.. pp. 183-244.
2000, "Ptoleinaic to Roman Egypt: The Death of the Dirigiste State?", in Lo
Cascio and Rathbone, pp. 44-55.
2001, "The 'Muziris' Papyrus (SB XVI11 13167): Financing Roman
Trade with India", in Alexandrian Studies II in Honour of Mostafa el
Abbadi, Bulletin de la Societe Archeologique d'Alexandrie 46, Alexandria,
pp. 39-50.
2003, "The Financing of Maritime Commerce in the Roman Empire, I-II AD",
in Lo Cascio 2003b, pp. 197-230.
Rauh, N. 1986, "Cicero's Business Friendships, Economics and Politics in the
Late Roman Republic", Aevum 60, pp. 3-30.
1989, "Auctioneers and the Roman Economy", Historia 38. pp. 451-471.
1993. The Sacred Bonds of Commerce. Religion, Economy, and Trade Society
at Hellenistic Roman Delos, 166-87 B. C., Amsterdam.
Ray, R. K. 1988, "The Bazaar: Changing Structural Characteristics of the Tndige-
nousSectionoftheIndianEconomy BeforeandAftertheGreatDepression",
Indian Economic and Social History Review 25, pp. 263-318.
Raychaudhuri, T. 1991, "The Commercial Entrepreneur in Pre-Cotonial India:
Aspirations and Expectations: A Note", in Ptak and Rothermund, pp. 339-
351.
Rea, J. 1982, "P. Lond. Inv. [562 Versa: Market Taxes in Oxyrhynchus",
Zeitschrift fur Papyrologieund Epigraphik46, pp. 191-209.
Reed. M. 1996, "London and its Hinterland 1600-1800", in P. dark and B.
Lepetit (eds. ), Capital Cities and their Hinterlands in Early Modern Europe,
Aldershot, pp. 51-83.
Reger, G. 1994, Regionalism and Change in the Economy of Independent Delos,
Berkeley, CA.
2003, "Aspects of the Role of Merchants in the Political Life of the Hellenistic
World", in Zaccagnini, pp. 165-198.
Reinhardt, F. 1968, Kapitalaufwendungen der Gesellschqfter und Gesellscha-
ftvermogen in der klassischen romischen Societas, Heidelberg.
Reinmuth, 0. W. 1936, "Two Prefecturial Edicts Concerning The Publican",
Classical Philology 31, pp. 146-162.
Remesal Rodrfguez, J. 1983, "Olproduktion und Olhandel in der Baetica",
Miinstersche Beitrdge wr Antiken Handelsgeschichte 2, 2, pp. 91-112.
1986, La annona militaris y la exportation de aceite betico a Germania,
Madrid.
i997> Heeresversorgiing und die Wirtschaftlichen Beziehungen wischen der
Baetica und Germanien, Stuttgart.
1998, "Baetican Olive Oil and the Roman Economy", in Keay, pp. 183-
199.
338
BIBLIOGRAPHY
2001, "Politik und Landwirtschaft im [mperium Romanum am Beispiel der
Baetica", in Peter Herz and Gerhard Waldherr (eds. ), Landwirtschaft im
Imperium Romanum, St Katharinen.
Ren.eer, J. 2003, "Trade and Market in the Ancient Near East: Theoretical and
Factual Implications", in Zaccagnini, pp. 15-39.
Reynolds, P. 1995, Trade in the Western Mediterranean. AD: 400-700. The
Ceramic Evidence, BAR International Series 600, Oxford.
Ricarclo, D. 1996, Principles of Political Economy and Taxation, Great Minds
Series. Amherst. NY.
Rich, J. and Shipley, G. (eds. ) 1993, War and Society in the Roman World,
London and New York.
Rich, J. and Wallace-Hadrill, A. (eds. ) 1991, City and Country in the Ancient
World. London and New York.
Richards, J. F. 1975, MughalAdministration in Golconda, Oxford.
1981, "Mughal State Finance and the Premodern World Economy", Compar-
ative Studies in Society andHistory 23, pp. 285-308.
1984, "Norms of Comportment among Imperial Mughal Officers", in B. D.
Metcalf (ed. ). Moral Conduct and Authority. The Place of Adab in South
Asian Islam, Berkeley, CA, pp. 255-289.
1986, Document Forms for Official Orders of Appointment in the Mughal
Empire, Cambridge.
(ed. ) 1987,The Imperial Monetary Systemof Mughal India, Delhi.
1990> "The Seventeenth Century Crisis in South Asia", Modern Asian Studies
24, 4, PP. 625-638.
i993a. The Mughal Empire, The New Cambridge History of India I, 5,
Cambridge.
i993b' Power, Administration, and Financein Mughal India, Aldershot.
Rickman, G. i98oa, The Corn Supply of Ancient Rome, Oxford.
i98ob, "The Grain Trade under the Roman Empire", in D'Arms and Kopff,
pp. 261-275.
Rizzo, G. 2003, Instrwnenta Urhis I. Ceramiche fini da mensa, lucerne ed anfore
a Roma nei primi due secoli dell'impero. Collection de 1'Ecole Fran^aise
de Rome 307, Rome.
Robert, L. 1969, OperaMinora Selecta, vol. I, Amsterdam.
Roberto, U. 2003, '"Del commercio dei Romani': Politica e storia antica nelle
riflessioni del Settecento", in Zaccagnini,pp. 327-361.
Robertson, J. 1997, "Gibbon's Roman Empire as a Universal Monarchy: The
Decline and Fall and the Imperial Idea in Early Modern Europe", in R.
McKitterick and R. Quinauldt (eds. ), Gibbon and Empire, Cambridge,
pp. 247-270.
Roda, S. 1998, "II Senato nell'alto impero Romano", in // Senate nella storia.
fnstituto Poligrafico e Zecca Dello Stato, Rome, pp. 129-221.
Rodger, A. 1996, "Postponed Business at Imi", Journal of Roman Studies 86,
pp. 61-73.
339
 BIBLIOGRAPHY
1997. "Vadimonium to Rome (and Elsewhere)", Zeitschrift der Savigny-
Stiftung. Romanistische Abfeilung 114, pp. 160-196.
Rodinson, M. 1966, Islamet capitalisme, Paris.
Rodnguez-Almeida, E. 1984, // Monte Testaccio. Ambiente, storia, materiali,
Rome.
Rostovtzeff, M. I. 1902, Geschichte der Staatspacht in der Romischen Kaiseneit
bis Diocletian, Philologiis, Suppl. 9, pp. 329-512.
19323, Caravan Cities, trans. D. and T. Talbot-Rice, Oxford.
l932b,"LesinscriptionscaravanieresdePalmyre",in MelangesGustaveGlotz,
vol. II, Paris, pp. 792-811.
i957>The SocialandEconomic History ofthe RomanEmpire,2ndedn, revised
by P. M. Fraser, Oxford.
Rothermunci, D. 1991, "Asian Emporia and European Bridge-Heads", in Ptak
and Rothermund, pp. 3-8.
Rotschild, M. 1973, "Models ofMarket Organization with Imperfect Information:
A Survey", Journalof Political Economy 81. pp. 1283-1308.
Rouge, J. 1966, Recherches sur I'organisation du commerce maritime en
Mediterranee soils I'empire remain, Paris.
1980, "Pret et societe maritime dans Ie monde romain", in D'Arms and Kopff,
pp. 291-303.
1981, Ships and Fleets of the Ancient Mediterranean, trans. Susan Frazer,
Middleton, CT.
1985, Droit romain et sources de richesses non foncieres", in Leveau 1985,
pp. 161-176.
Royden, H. L. 1988, The Magistrates of the Roman Professional Collegia in
Italy. From the First to the Third Century A. D., Pisa.
Ruggini, L. C. 1973, "Stato e associazioni professionali nell'eta imperiale
romana in Akten des VI. Internationalen Kongresses fur Griechische imd
Lateinische Epigraphik, Munich, pp. 271-311.
Ruprechtsberger, E. M. (ed. ) 1987, Palmyra. Geschichte, Kunst und Kultur der
Syrischen Oasenstadt, Linz.
Sabloff, J. A. and Lamberg-Karlowsky, C. C. (eds. ) 1975, Ancient Civiliwtion
and Trade, Albuquerque, NM.
Sahlins, M. 1972, Stone Age Economics, Chicago.
I995> ^iow "Natives" Think About Captain Cook For example, Chicago and
London.
2000, "Cosmologies of Capitalism: The Trans-Pacific Sector of 'The World
System'", in M. Sahlins, Culture in Practice. Selected Essays, New York,
pp. 415-469.
Said, E. 1978, Orientalism. Western Conceptions of the Orient, New York and
London.
1993, Culture and Imperialism, London.
Sallares, R. 2002, Malaria and Rome. A History of Malaria in Ancient Italy,
Oxford.
340
BIBLIOGRAPHY
Saller. R. 1982, Personal Patronage under the Early Empire, Cambridge.
i994t Patriarchy, Property and the Roman Family, Cambridge.
2000, "Statusand Patronage",in The CambridgeAncient History, vol. XI, 2nd
edn, Cambridge, pp. 817-854.
2002, "Framing the Debate over Growth in the Ancient Economy , in Scheidel
and von Reden. pp. 25 1-269. (Also published in Morris and Manning 2005,
pp. 223-238.)
Santucci, G. 1997, // socio d'opera in diritto romano. Conferimenti e respons-
abilita. Padua.
Salmeri, G. 1998, "Per una lettura dei capitoli V-VII Della Storia economica e
sociale dell'Impero Romano di M. Rostovtzeff, Athenaeum 86, pp. 57-84.
Sarris, P. 2006, Economy and Society in the Age ofJustinian,Cambridge.
Sartre, J. -P. 1961, "Preface", in Franz Fanon, Les Damnes de la terre, Paris.
Sartre, M. 1991, L 'Orient romain. Provinces et societes provinciales en
Mediterranee orientale d'Auguste aux Severes (31. avant J.C. -2J5 apres
J. C. ), Paris.
Scheidel, W. 2000, Debciting Roman Demography, Leiden.
2002, "A Model of Demographic and Economic Change in Roman Egypt after
the Antonine Plague", Journalof Roman Archaeology 15, pp. 97-114-
2007a, "Demography", in I. Morris, R. Saller and W. Scheidel (eds. ). The
Cambridge Economic History of the Greco-Roman World, Cambridge,
pp. 38-86.
2007b, "A Model of Real Income Growth in Roman Italy", Historia 56, pp.
332-346.
(ed. ) 2008, Rome and China: Comparative Perspectives on Ancient World
Empires, New York.
Scheidel. W. and von Reden, S. 2002, The Ancient Economy, Edinburgh.
Schiavone, Aldo 2000, The Endof the Past, Cambridge, MA.
Schleich, T. 1983, "Uberlegungen zum Problem senatorischer Handelsak-
tivitaten. Teil I. Senatorische Wirtschaftsmentalitat in modemer und antiker
Deutung", Miinstersche Beitrdge wr antiken Handelsgeschichte i, 2,
pp. 65-90.
[984, "Oberlegungen zum Problem senatorischer Handelsaktivitaten. Teil
2. Zwischen 'otium' und 'negotium': Gelegenheitsunternehmungen und
domestizierte Wirtschaft", Miinstersche Beitrdge wr antiken Handels-
geschichte 3, [, pp. 37-76.
Schmidt-Colinet, A., Stauffer, A. and Al-A.s'ad, Khaled 2000, Die Textilien aus
Palmyra, Mainz.
Schneider, H. 1990, "Die Biicher-Meyer Kontroverse", in Calder and Demandt,
pp. 417-445.
Schneider, Jane. 1977, "Was there a Pre-capitalist World System?", Peasant
Studies 6, I, pp. 30-39.
Schneider, Jiirgen 1986, "The Significance of Large Fairs, Money Markets and
Precious Metals in the Evolution of a World Market from the Middle Ages to
341
 BIBLIOGRAPHY
the First HalfoftheNineteenthCentury", in Fischer,MclnnisandSchneider,
pp. 15-36.
Schwarz, H. 2001a, "Anmerkungen zu der Zollinschrift aus Myra", Epigraphica
Anatolica 33, pp. 15-37.
2001 b, Die Finanwirtschaft kleinasiatischer Stdclte in der romischen
Kaiserwt, Bonn.
Schwarz,L. D. 1992, Londonin the Age oflndustricilization,Cambridge.
Scullard, H. H. 1980, A History of the Roman World 753 to 146, 4th edn, London.
Selancl,T. 1996."PhiloandtheClubs andAssociationsofAlexandria",in Klop-
penborg and Wilson, pp. I [0-127.
Serrao, F. 2000, "Impresa, mercato, diritto: riflessioni minime", in Lo Cascio
2000C, pp. 31-68.
Shanin,T. (ed. ) 1987, Peasantsand Peasant Societies, 2nd edn, Oxford.
Sharma, R. S. 1970, "The Ardha-Kathanak: A Neglected Source of Mughal
History", Indica7, pp. 49-73 and 106-120.
Sharp, M. 1998, The Food Supply in Roman Egypt, D. Phil. thesis. University of
Oxford.
Shatzman, I. 1975, Senatorial Wealth and Roman Politics, Brussels.
Shaw,B. D. 1981,"RuralMarketsin North AfricaandthePoliticalEconomyof
the Roman Empire", Antiquitates Africanae 17, pp. 37-83.
1984, "Bandits in the Roman Empire", Past and Present 105, pp. 3-52.
2001, "Challenging Braudel: A New Vision of the Mediterranean", Journal of
Roman Archaeology 14, pp. 1-25.
Sherk, R. K. 1955, "The Inermes Provinciae of Asia Minor", American Journal
ofPhilology 76, pp. 400-413.
1957, "Roman Imperial Troops in Macedonia and Achaea", American Journal
ofPhilology 78, pp. 52-62.
(ed. and trans. ) 1988, The Roman Empire. Augustus to Hadrian, Cambridge.
Sherwin-White,A. N. 19730,The Roman Citizenship, 2nd edn, Oxford.
I973b, The Tabula of Banasa and the Constitutio Antoniniana", Journal of
Roman Studies 63, pp. 86-98.
Sijpestein, P. J. 1987, Customs Duties in Graeco-Roman Egypt. Studia Amster-
damensia ad Epigraphicam, lus Antiquum et Papyrologicam Pertinentia 17,
Zutphen.
Singh, C. 1988, "Centre and Periphery in the Mughal State: The Case of
Seventeenth-century Panjab", Modern Asian Studies 22, pp. 299-318.
Singh, M. P. 1972, "Merchants and the Local Administration and Civic Life in
Gujarat During the [ 7th Century", in Medieval India. A Miscellany, vol. 2,
pp. 221-226, n. p.
1985, Town, Market, Mint and Port in the Mughal Empire, 1556-1707, New
Delhi.
Sirks, B. 1991, Food for Rome, Amsterdam.
2002, "Sailing in the Off-Season with Reduced Financial Risk", in Aubert and
Sirks, pp. 134-150.
342
BIBLIOGRAPHY
Skinner, G. W. 1964, "Marketing and Social Structure in Rural China", Journal
of Asian Studies 24, I, pp. 3-43; 24, 2, pp. 195-228; and 24, 3. pp. 363-
399.
Skocpol, T. 1979, States and Social Revolutions. A Comparative Analysis of
France, Russia and China, Cambridge.
Skocpol, T. and Sommers, M. 1980, "The Uses of Comparative Historical
Inquiry", Comparative Studies in Society and History 22, pp. 174-194.
Skydsgaard,J. E. 1976, "The Disintegrationof the Roman LabourMarket and
the Clientela Theory", in K. Ascani et al., Studia Romana in Honorem Petri
Krarup Septuagenarii,Odense, pp. 44-48.
i997> "Theodor Mommsens romerske kejserhistorie", Historisk Tidsskrift 97,
2, pp. 305-3I 2.
Smith, Adam 1976, An Inquiry' into the Nature and Causes of the Wealth of
Nations. Oxford.
Smith, Anthony, D. 1986, The Ethnic Origin of Nations, London.
Snell, R. 2005, "Confessions of a 17th-century Jain Merchant: The Ar-
dhakathanak of Banarasfdas", South Asia Research 25, pp. 79-104.
Sombart, W. 1913, Luxusund Kapitcilismus. Leipzig.
Sommer, M. 2005, Roms orientalische Steppengreiw. Palmyra-Edessa-Dura-
Eiiropos-Hatra. Eine Kiilturgeschichte van Pompeius bis Diocletian, Wies-
baden.
S0rensen, G. 1994, "Hvad er udvikling?", Den Jyske Historiker 66, pp. 19-32.
Steensgaard, N. 19730, Carracks, Caravans and Companies. The Structural
Crisis in the European Asian Trade in the Early i-jth Century, Odense.
l973b, "Universal History for Our Times", Journal of Modem Histoi-y 45,
pp. 72-82.
1980, "Asian Trade i5th-i8th Centuries: Continuity and Discontinuities",
XVe Congres International des Sciences Historu/ues Rapports II, Bucharest,
pp. 488-500.
1981, "Violence and the Rice of Capitalism: Frederic C. Lane's Theory of
Protection and Tribute", Review 5, 2, pp. 247-273.
1984, "Set t'ra 1984. En model for nyere tids verdenshistorie", in Tradition
0,1; Kritik. Festskrift til Svencl Elleh0j den 8. September 1984, Copenhagen,
pp. 413-435.
1987, "The Indian Ocean Network and the Emerging World-Economy
(c. 1550-1750)", in S. Chandra (ed. ). The Indian Ocean. Explorations in
Hlston, Commerce, and Politics, New Delhi, pp. 125-150.
i990a, "The Growth and Composition of the Long-distance Trade of England
and the Dutch Republic before 1750", in Tracy, pp. 102-152.
l99°b, "Opdagelsemes plads i verdenshistorien. Varer, ^delmetal og Tjene-
steydelser i International Handel f0r 1750", Historisk Tidsskrift 90,
pp. 221-246.
I990C, "The Seventeenth-Century Crisis and the Unity ot'Eurasian History ,
Modern Asian Studies 24, 4, pp. 683-697.
343
 BIBLIOGRAPHY
1991. "Emporia: Some Reflections", in Ptak and Rothermund, pp. 9-12.
Stein, A. 1927, Der romische Ritterstand. Ein Beitrag wr social- und personen-
ffeschichte des romischens Reiches, Munich.
Stein, B. 19853, "State Formation and the Economy Reconsidered. Part I",
Modern Asian Studies 19, 3, pp. 387-413.
i985b, "Politics, Peasants and the Deconstruction of Feudalism in Medieval
India", in Byres and Mukhia, pp. 54-86.
1998. A History of India, Oxford.
Stigler, G. J. and Sherwin, R. A. 1985, "The Extent of the Market", Journal of
Law and Economics 28, pp. 554-577.
Stocking G. W., Jr 1995, After Tylor. British Social Anthropology TSSS-rgyr,
London.
Stoianovich, T. 1974, "Pour un modele du commerce du Levant: economie con-
currentielle et economie de Bazar 1500-1800", Bulletin. Association Inter-
nationale d'Eludes du Sud-Est Europeen 12, 2, pp. 6[-120.
1994, "Cities, Capital Accumulation, and the Ottoman Balkan Command
Economy, 1500-1800", in Tilly and Blochmans, pp. 60-99.
Stuart, P. and Bogaers, J. E. (eds. ) 1971, Deae Nehalenniae. Gids Bij De Ten-
toonstelling Mehalennia de Zeeuwse Godin, Leiden.
2001, Nehalennia. Romische Steindenkmdler aus der Oosterschelde hei Co-
lijnsplaat, Leiden.
Subrahmaniam, L. 1999, "India's International Economy, 1500-1800", Indian
Historical Review 25, 2, pp. 38-57.
Subrahmanyam,S. 1989,"StateFormationandTransformationin Early Modem
India and Southeast Asia", in P. J. Marshall etal. (eds. ) India and Indonesia
during the Ancien Regime, Leiden, pp. 91-109.
(ed. ) 19900, Merchants, Markets and the State in Early Modem India. Delhi.
i99ob.The Political Economyof Commerce, Cambridge.
(ed. ) 1994, Money and the Market in India i loo-f/oo, Delhi.
1995."OflmaratandTijarat: AsianMerchantsandStatePowerin theWestern
Indian Ocean, 1400 to 1750", Comparative Studies in Society and History
37, PP. 750-780.
(ed.) 1996, Merchant Networks in the Early Modern World, An Expanding
World 8, Aldershot.
I998' "Hearing Voices: Vignettes of Early Modernity in South Asia, 1400-
l750"> Daedalus 127, 3, pp. 75-104.
Swain, S. 1996, Hellenism and Empire. Language, Classicism, and Power in the
GreekWorld, AD50-250, Oxford.
Syme, R. 1958, Tacitus, vols. 1-2, Oxford.
1977. La richesse des aristocraties Betique et de Narbonnaise", Ktema 2,
pp. 373-380.
Talamanca, M. 1955, "Contributi allo studio delle vendite all'asta nel mondo
classico", Atti della Accademia Nawmale dei Lincei 6, pp. 35-251.
Talbert, R. J. A. 1984, The Senate in Imperial Rome, Princeton, NJ.
344
BIBLIOGRAPHY
Tandrup, L. 1979, Ravn, vols. 1-2, Viborg.
Tate, G. 1992, Les Campagnes de la Syrie du nord du lie au Vile siecle. Un
exemple d expansiondemographu/ue et economique a la fin de I'antiquite,
Paris.
Taubenschlag,R. 1955,The Lawof'Greco-RomanEgyptinthe Lightofthe Papyri
(332 B. C. -A. D. 640), 2nd edn, Warsaw.
Tavernier, J. B. 1889, Tavernier's Travels in India, vols. 1-2, ed. and trans.
V. Ball, London (original French edn 1676).
Tchemia, A. 1983,"ItalianWinein Gaul at the Endofthe Republic", in Gamsey,
Hopkins and Whittaker, pp. 87-102.
1986, Le Vin de I'ltalie romaine, Paris.
1989, "Encore sur les modeles economiques et les amphores", in Amphores
romaines et histoire economique. Dix ans de recherche. Collection de
i'Ecole Fran^aisede Rome I 14, Rome, pp. 529-536.
1992, "Le dromadaire des Peticii et Ie commerce Oriental", Melanges de
I'Ecole Frangaisede Rome. Antiqiiite 104, pp. 293-301.
1995, "Moussons et monnaies: les voies du commerce entre Ie monde Greco-
Romain et 1'Inde", Annales. Histoire, Sciences Sociales 50, pp. 991-1009.
2000, "Lavente du vin", in Lo Cascio 20000, pp. 199-211.
(ed. ) 2005, Autour de la rationalite antique, in Topoi. Orient-Occident 12-13,
pp. 259-314.
2006, "Lacrise de L'ltalie imperialeet la concurrencedesprovinces", Cahiers
du Centre de Recherches Historiques 37, pp. 137-156.
Teixiclor,J. 1984, Un Port romaindu desert. Palmyre et soncommerce d'Auguste
a Caracalla, Semitica 34, Paris.
Temin, P. 2001, "A Market Economy in the Early Roman Empire", Journal of
Roman Studies 91, pp. 169-181.
2006, "Estimating GDP in the Early Roman Empire", in E. Lo Cascio (ed. ),
Innovawme tecnica e progresso economico nel mondo romano. Ban,
PP. 31-54.
Temin, P. andKessler, D. 2007 "The Organizationofthe GrainTrade in the Early
Roman Empire", Economic History Review 60, 2, pp. 313-332.
Thebert, Y. 1980, "Economie, societe et politique aux deux demieres siecles de
la republique romaine", Annales. Histoire, Sciences Sociales,pp. 875-911.
Thompson, E. P. 1993, Customs in Common. Studies in Traditional Popular
Culture, New York.
Thrupp, S. L. 1958, "Editorial", Comparative Studies in Society and History I,
PP. 1-4.
Tilly, C. 1984, Big Structures, Large Processes, Huge Comparisons, New York.
I992' Coercion, Capital and European States, A. D. 990-7992, Oxford.
Tilly, C. and Blockmans, W. (eds. ) 1994, Cities and the Rise of States in Europe,
A. D. 1000-1800, Boulder, CO.
Tomber,R. 1993,"QuantitativeApproachesto theInvestigationofLong-Distance
Exchange",Journal of Roman Archaeology 6, pp. 142-166.
345
 BIBLIOGRAPHY
Toner, J. P. 1995, Leisure and Ancient Rome. Cambridge.
Torri, M. 1998,"MughalNobles, IndianMerchants, and the Beginningof British
Conquest in Western India: The Case of Surat 1756-59", Modern Asian
Studies 32. 2. pp. 257-315.
Tracy, J. D. (ed. ) 1990, The Rise of Merchant Empires. Long-DistanceTrade in
the Earlv Modern World. Cambridse.
(ed. ) [991, The Political Economy of Merchant Empires, Cambridge.
Treadgold, W. 1997, A History ofthe Byzantine State and Society, Stanford, CA.
Tylor, E. B. 1871, Primitive Culture, London.
Udovitch, A. L. 1962, "At the Origins of the Western Comm. enda: Islam, Israel,
Byzantium", Speculum 37, pp. 198-207.
1970. Partnership and Profit in Medieval Islam, Princeton, NJ.
Veblen, T. 1899, The Theory of the Leisure Class, New York.
Velissaropoulos, J. 1980, Les Naucleres grecs, Paris.
Vera, D. 1994, "L'ltalia agraria nell'eta imperiale: Fra crisi e transformazione",
in L'ftalie d'Aujfuste a Diocletien, Collection de 1'Ecole Fran^aise de Rome
198, Rome, pp. 239-248.
Verboren, K. 2002, The Eccmomy of Friends. Economic Aspects of "Amicitia"
and Patronage in the Late Republic. Collection Latomus 269, Brussels.
Vemhet, A. 1991, La Graufesenque ceramiques gallo-romaines, Millau.
Veyne, P. 1961, "Vie de Trimalchion", Annales ESC 16, pp. 213-247.
1971, Comment on ecrit I'histoire. Essai d'epistemologie, Paris.
i97^a L'lnventaire des differences, Paris.
i976b, Le Pain et Ie cirque, Paris. (Published in English, 1990.)
1979. "Rome devant la pretendue t'uite de I'or: mercantilisme ou politique
disciplinaire?", Annales ESC, pp. 21 1-241.
i990> Bread and Circuses, trans. B. Pearce, London.
Virlouvet, C. (ed. ) 1997, La Rome imperiale. Demographie etlogistique, Rome.
Visscher, F. de 1940, Les edits d'Auguste decouverts a Cyrene, Louvain.
Vittinghof,F. ic)^, "PoTtoTmm", Paulys RealencyclopddiederclassischenAlter-
tumswissenschaft, 22, i, cols. 346-399.
Voltaire, P. M. A. de 1963, Essai sur les moeurs et I'esprit de nations et sur les
principaux fails de I 'histoire depuis Charlemagne jusqii 'a Louis XIII, 2 vols.,
Paris.
Vries,J.de 1976, EconomyofEurope inan AgeofCrisis , 600-7750,Cambridge.
1984, European Urbanization, 1500-1800, London.
1990; "Problems in the Measurement, Description, and Analysis of Historical
Urbanization", in A. Woude, A. Hayomi and J. de Vries (eds. ), Urbcin-
ization in History. A Process of Dynamic Interactions, Oxford, pp. 43-
6o.
Vries, J. de and Woude, A. 1997, The First Modern Economy. Success, Failure
and Perseverance of the Dutch Economy, 1500-1800, Cambridge.
Wacke, A. 1993, "Gallisch, Punisch, Syrisch oder Griechisch statt latein? Zur
Schrittweisen Gleichberechtigung der Geschaftssprachen im romischen
346
BIBLIOGRAPHY
Reich", Zeitschrift der Savigny-Stiftung fur Rechtsgeschichte. Romanistis-
che Abteilung 110, pp. 14-59.
1994. "Die adjektivischen Klagen im Uberblick, I: Von der Reeder- und der
Betriebsleiterklage zur direkten Stellvertretung", Zeitschrift der Savigny-
Snftungfur Rechtsgeschichte. Romanistische Ahteilung I[I, pp. 280-362.
Wagner, D. 2001, "The Administration of the Iron Industry in Eleventh Century
China", Journal of the Social and Economic History of the Orient 44, 2,
pp. 175-197.
Wallace-Hadrill, A. (ed. ) 1989, Patronage in Ancient Society, London and New
York.
1991. "Elites and Trade in the Roman Town", in Rich and Wallace-Hadrill
1991. PP. 241-272.
1994, Houses and Society in Pompeii and Herculaneum, Princeton, NJ.
Wallerstein, I. 1974;!, The Modern World-System, vol. I, San Diego, CA.
l974b, "TheRiseandFutureDemiseofthe WorldCapitalistSystem: Concepts
for Comparative Analysis", Comparative Studies in Society andHistory 16,
pp. 387-415.
1979. "The Ottoman Empire and the Capitalist World-Economy: Some Ques-
tions for Research", Review 2, 3, pp. 389-398.
Waltzing,J. P. 1895-1900,Etude historique sur les corporationsprofessionelles
chez. les Roinains, vols. 1-4, Louvain.
Watson, A. 1992.The Evolutionof InternationalSociety. A Comparative Histor-
ical Analvsis. London.
Weber, M. 19200, "Die protestantische Ethik und der Geist des Kapitalismus",
Gesammelte Aufsatze zur Religionssoziologie,vol. I, Ttibingen.
i920b, 'Vorwort' in M. Weber, Gesammelte Aufsdtz. e zur Religionssoziologie
vol. I, Tiibingen.
1922, Gesammelte Aufsiitz.e zur Wissenschaftslehre,Tiibingen.
l924a: "Agrarverhaltnisse im Altertum", Gesammelte Aufsdtz. e wrSozial- und
Wirtschaftsgeschichte, pp. 1-288, Tubingen.
i924b: "Die sozialen Grunde des Untergangs der antiken Kultur", Gesammelte
Aufsdtzezur Sozial-undWirtschaftsgeschichte,pp. 289-311,Tubingen.
19240: "Zur Geschichte der Handelsgesellschaften im Mittelalter", Gesam-
melte Aufsdtze wr Social- und Wirtschaftsgeschichte, Tiibingen, pp. 312-
443.
1972, Wirtschaft und Gesellschaft, Tubingen.
Wee, C. J. 1997, "Framingthe New East Asia", in R. Salim (ed. ), "The Clashof
Civilizations?" Asian Responses, Karachi, Oxford and New York, chapter
5.
Weeber, K. -W. 2003, Luxus im alten Rom. Die Schwelgerei, das siisse Gift,
Darmstadt.
2006, Liixus im alien Rom. Die dffentliche Pracht, Darmstadt.
Whitbread, I. K. 1995, Greek Transport Amphorae. A Petrological andArchae-
ological Study, Athens.
347
 BIBLIOGRAPHY
Whittaker, C. R. 1985, "Tradeand the Aristocracy in the Roman Empire", Opus,
International Journal for Social and Economic History of Antiquity 4, pp.
49-75.
1989, "Amphorae and Trade", in Amphores romaines et histoire economiqiie.
Dixans de recherche. Collection de 1'Ecole Fran^aisede Rome 114, Rome,
PP. 537-539.
i993> Land, City andTrade in the Roman Empire, Aldershot.
i994a> Frontiers of the Roman Empire. A Social, and Economic Study, Balti-
more. MD.
i994b, "The Politics of Power: The Cities of Italy", in L'ltalie d'Auguste a
Diocletien, Collection de 1'Ecole Fran^aise de Rome 198, Rome, pp. 127-
i43.
l995> "Do Theories of the Ancient City Matter?", in Cornell and Lomas,
pp. 9-26.
2002, "Proto-industrialization in Roman Gaul", in Karen Ascani et al. (eds. ),
Ancient History Matters. Studies Presentedto Jens Erik Skydsgaardon His
SeventiethBirthday, Analecta RomanaInstitutiDanielSuppl. 30, pp. l 1-22.
Wickham, C. 1984, "The Other Transition: From the Ancient World to Feudal-
ism". Past mul Present 103, pp. 3-36.
1985, "The Uniqueness of the East", Journal of Peasant Studies 12, pp. 167-
196.
1988, "Marx, Sherlock Hoknes and Late Roman Commerce", Journal of
Roman Studies 78, pp. 183-193.
2005, Framing the Early Middle Ages, Oxford.
Wierschowski, L. 1995, Die Regionale Mobilitat in Gallien nach dem Inschriften
des f. -3. Jahrhunderts n. Chr., Historia Einzelschriften 91, Stuttgart.
Will, E. 1957, "Marchandset chefs de caravannesa Palmyre", Syria34, pp. 262-
277.
1992, Les Palmyreniens. LaVenise des sables (ler siecle avant- IIIemesiecle
apres J. C. ), Paris.
Wills, J. E., Jr. 1993, "Maritime Asia, 1500-1800: The Interactive Emergence of
European Domination", American Historical Review 98, pp. 83-105.
Wilson, A. 2001, "Timgad and Textile Production", in Mattingly and Salmon,
pp. 271-296.
2002,"Machines,PowerandtheAncientEconomy",JournalofRomanStudies
92, pp. 1-32.
Wink, A. 1989, " 'Al-Hind' India and Indonesiain the Islamic World-Economy,
ca. 700-18oo",in P.J.Marshalletal., IndiaandIndonesiaduringthe Ancien
Regime, Leiden, pp. 33-72.
Wittfogel,K. A. 1957,OrientalDespotism: A ComparativeStudy ofTotalPower,
New Haven, CT.
Wolf, E. 1966, Peasants, Englewood Cliffs, NJ.
1982, Europe and the People Without History, Berkeley and Los Angeles.
Wolff, C. 2003, Les Brigands en orient sous Ie haut-empire romain, Rome.
348
BIBLIOGRAPHY
Wolff, H. J. etal. 1978, Handhuch der Altertumswissenschaft, 10. 5. 2: Das Recht
dergriechischenPapyriAgyptens in cierZeitderPtolemaeerund des Prinzi-
pats, Munich.
1980, "Romisches Provinzialrecht in der Provinz Arabia", Aufstieg imdNieder-
gang der RomischenWelt 2, 13, pp. 763-806.
Walters, R. 1999, Nummi Signati. Untersiichungen zur rdmischen Mimzpragung
und Geldwirtschaft, Munich.
Wong, R. B. 1997, ChinaTransformed, Ithaca, NY.
Woolf, G. 1990, "World-systems Analysis and the Roman Empire", Journal of
Roman Archaeology 6, pp. 44-58.
1992, "Imperialism, Empire and the Integration of the Roman Economy",
World Archaeology 23, 3, pp. 282-293.
1993. "RomanPeace", in Rich and Shipley, pp. 171-194.
1997. "The Roman Urbanization of the East", in Alcock 19970, pp. 1-14.
1998, Becoming Roman. The Origins of Provincial Civilimtion in Gaiil, Cam-
bridge.
2001,"RegionalProductionsin Early RomanGaul", in Mattingly and Salmon,
pp. 49-65.
Wrigley, E. A. 1987, Peoples, Cities and Wealth. The Transformation ofTradi-
tional Society. Oxford.
1988, Continuity, Chance and Change. The Character of the Industrial Revo-
lution in England, Cambridge.
Yang, A. A. 1998, Bazaar India. Markets, Society, and the Colonial State in
Gangetic Bihar, Berkeley, Los Angeles and London.
Yaron, R. [964 "Reichsrecht, Volksrecht and Talmud", Revue Internationale des
Droits de I'Antiquite, yd ser. 2, pp. 281-298.
Yon. J.-B. 2002. Les Notables de Palrnvre. Beirut.
Young, G. 2001, Rome's Eastern Trade. International Commerce and Imperial
Policy, 31 BC-AD 305, London.
Zaccagnini,C. (ed. ) 2003, Mercanti e politica net mondo antico, Rome.
Zanker, P. 1987, Augustus und die Macht der Bilder, Munich (English
translation by A. Shapiro, Ann Arbor, MI, 1988.)
Zelener, Yan 2006, "Between Technology and Productivity", in E. Lo Cascio
(ed. ), Innowione tecnica e progresso economico net mondo romano. Ban,
pp. 303-18.
Zevi, F. 1989. "Introduzione", in Amplwres romaines et histoire economique.
Dix ans de. recherche. Collection de I'Ecole Fran^aise de Rome 114, Rome,
pp. 3-19.
Ziche, H. 2006, "Integrating Late Roman Cities, Countryside and Trade", in
Bang, Ikeguchi and Ziche, pp. 255-276.
Ziegler, K. -H. 1971, Das private Schiedsgericht im antiken romischen Recht,
Munich.
Zimmermann, R. 1990, The Law of Obligations. Roman Foundations of the
Civilian Tradition, Cape Town.
349
 BIBLIOGRAPHY

0rsted, P. 1985, Roman Imperial Economy and Romanization, Copenhagen.

2000, "Roman State Intervention? The Case of Mining in the Roman Empire",
in Lo Cascio and Rathbone. pp. 70-76.
INDEX
0rsted, P., Carlsen, J., Ladjimi Sebai, L. and Ben Hassen, H. (eds. ) 2000, Africa
Proconsularis. RegionalStudiesinthe SegennesValley ofNorthernTunisia,
vol. 3: Historical Conclusions, Aarhus.
0stergard, U. 1973, "Den Antikke Procluktionsmade. Oversigt over histo-
rieskrivningen om forholdet mellem 0konomi og samfund i Antikken", Den
Jyske Historiker 5-6, pp. 4-42. AbuaI-Fazl 83 archaeology 2-3, 35, 288, 304
1976. "Studiet af ferkapitalistiske samfundsforhold - Mulighed og relevans. adulteration of goods 194 and economic integration 74, 287-9
Teoretiske og metodiske problemer". Den Jyske Historiker 6. pp. 52- Africa 2, 73, 77 difference in scale between Greek and

135. agency see business managers Roman economy 114

1996, "The Meaning of Europe - Empire, Nation-states, Civilization", Agra 9. 145. >99 Arezzo 288

Arhejdspapirer fra Center for Kulturforskning, Aarhiis Vniversitet 29, agrarianate society 11, 43-5, 85 see also terra sif;illata
Aarhus. status and contract 18-19 Aristeides, Publius Aelius 290-1
agricultural economy, limits of i 16-19 aristocratic investment in trade 270,
1997' "Hvad nu hvis. . . Kontrafaktiske hypoteser og "abne" situationer i his-
agricultural surplus, disposable 282-4
tone og historiefbrskning", Arbejdspapirerfra Centerfor Kulturforskning, commercialisation of [ 19-2 l Arles 247
Aarhus Universitet 46, Aarhus. quantification ot 91-3 army, Roman and Mughal 94
and the state 84-5 tasks of the Roman 233-4, 240
see also agriculture and trade Asian Customs Law 204-38 passim
agriculture 6 A.skalon 273
and trade 22-4. 50-2, 77-9 Assyrian merchants 30
Akbar 80 see also Mesopotamia, ancient
Akerlof, G. 139-40 Attalid kingdom 221
Aleppo 305 Augsburg 55
Alexander Severus 277 Aueusta Troas 216
Alexandria 77, 126, 142, 177-9, 228-9, Augustine 247
238, 245, 273. 281-305 Augustus 216
amphorae 35, 73. 74, 194-5, 273 Aurangzeb 80, 215, 244, 245
Amsterdam 147 Aztec Empire 7
Anatolia 227
Annales historians 34 Babatha archive 185-6
annona 247 Babur 79
see also redistribution Babylon 154, 254
anthropology 6, 24-5, 35> 37 Baltic grain trade 147
see also bazaar: Geertz; Malinowski; Banarasidas 150-1, 239, 243, 250, 260-1,
"otherness"; Polanyi; Sahlins 269
Antioch 126, 142, 177, 179 bandits 226
famine of 143-4 Bayly, C. A. 257, 302
Antonine emperors 61 bazaar 1-2, 6, 12, 129, 131, 294-6
Antonine Plague 86, 123-5, asymmetries and imbalances of 149-50
173 commercial strategies of 199-289
Antoninus Pius 108 passim
Apuleius 137, l5l community 240-68, 296-8
Arabia 202 definition of 4, 197-201
Arcadia (late Roman province in imperial styles of
Egypt) 157-66 consumption 294-301, 306

350 351
 INDEX INDEX

bazaar (conf.) concessions obtained from Mughal Corrupting Sea, The 27, fd-l, 136-7 market integration in 153-73
local monopolies of 139-40, 143-6, emperor by 213-14 see also connectivity; ecology; Pharaonic 19
200 coordination of markets 145-6 primitivism-modernism Eisenstadt, S. N. 65
as a social system 197-8, 241 organisation of 145-9 credit and banks Elaea 216
Beijing 26-8, 66 Chayanov, A. V. 78 bottomry loans 270, 28o-l Eleusis 151
Bengal 9 Chinese Empire 51-4, 62 Indian 268-9 elite, Roman imperial 96-1 lo
Bervtus 280 chtysargyrcm 226, 233 short-term 269, 272-3 incomes of 111-13
bishops 263 Cicero, Marcus Tullius 106, 239, 241 "crisis of Italy" 74 empire
Black Death 123, 173 city states 104 customs dues [60-2, 202-"i8, colonial 10, 40, 243
see also Antonine Plague Italian 271 295-8 early modern Asian 7-10, 12, 25. 43,
Bloch, Marc 9 classics 6 concessions 227-9 54-60, 8o, 305
Bosporus, Strait of the 221-2 clientelisation 5. 201, 287-9 declaration of goods 236 failure of in Europe 55
Bouclica 101 see also credit; fides local and state 220-5 patrimonial-bureaucratic 64, 82 see
Braudel, Fernand 46, 50, 86, J 34 collegia 156-7, 242, 262, 267-8 multiple payment of 218-20 also empire, tributary; Weber
Bronze Age civilisation 44 ordinances of 263-4 political favour and preference 215-16 pre-Columbian 7
see also agrarianate society potential for social mobilization 244-5 and surcharges 205-6 tributary 10-11, 59-6o, 62, 122, 132,
Brundisium 280 patrons ol 249 see also protection 149
Bubon 232 see also communal association Cyprus 192 as agrarian empire io
Biicher, Karl 19-21, 23, 38 Columella 32 as concentrated mass 125
bureaucracy 65 commenda 270-2, 282 Da Gama, Vasco 41 and economic integration 110-21
business managers 276-80 see also business managers,peculium Damascus 2 and European state-system 54-60
see also commenda, pecuiiiim commercial customs and practices, Danzig 147 and institutional integration r 79
Byzantium 240 regional variation of 190-5, de-colonisation 24, 40 and modernisation 65-6
295-9 see also Orient/Orientalism; and parasitical fiscalism 21 3
cahotage 141-2 commercial imperialism, Portuguese 306 "otherness' as revenue pump 95, 109
Caesarea(in Cappaclocia) [40 Commodus 6l, 2-i2 Delhi 66, 80 as universal empire lo
Cairo 43, 189, 3"5 communal associations 241-68 Red Fort of 80 see also Aztec Empire; Chinese Empire;
capital and cargoes, parcelling of [95, and medieval guilds 241-2, 248 Delos so IncaEmpire; Mughal Empire;
269-70, 273, 275 collective action 248-9 Denderah 254 Parthian Empire
capitalism 2, 4-5> 8, 32, 45-6. 271, 291 corporate bodies 246-8 desert, Egyptian 23 [ Ephesos 77, 142, 178, 2l6
and the bazaar 150, 198, 295-8 government suspicions against 245-6 despotism 53, 61, 63, 96, 292-3 bakers in 248
and tributary empire 11, 235 see also collegia differences,cultural 9 Epiphanius 192
see also Europe, early modem compared comparative history 7-10, 24-5, 292 Dio Chrysostomus 77 eurocentnsm
to Rome; mercantilism and the notion of functional Dionysia, trial of 188 and global history 38-45
capstone government 66, [ l I equivalent 35-6, 49 Domitian 98, 103 Europe, early modern compared to
Caracalla 98, 184 conceptualisation 3-6, 26-8, 115 Droysen, Johann Gustav 3-4 Rome 33-6, 47, 71, 292
caravans 134, 231, 258 connectivity 71, 137 Dumont, Louis 35, 242 see also capitalism, mercantilism
Carthage 126, 142 Constantinople 57, 247 Dura Europos 19 [, 254 evidence 2-3, 222-3, 248, 250, 263
Cassiodorus 138 Constitutio Antoniniana 184 astronomical tablets of Hellenistic
Cassius Dio 98 consumer city/interests 9, 27-8, 47-8, 67, East IndiaCompany (English) Babylon 154
caste 242-4, 262 152-3, 213, 295-8 and markets in Bengal I 44 the Egyptianprice records 153-7, 294,
and medieval guilds 244 consumption, imperial styles of 297-301, ecology and economic integration 70-2 301, 306
Cato 120 304 economics, discipline of 6-7 exchange, levels of interregional I 13
charismatic goods 302-3 conspicuous 295-9 economy, moral 152-3, 254
see also consumption regulated by fashion 299 see also consumer city/interests fairs, periodic 49-50, 197-8
Charlemagne 55 Cook, James, death on Hawaii 37-8 economy, quantifying size of 86 falling marginal returns 123-4
Charles V (Holy Roman Emperor) 55 Coptos 174, 254, 281 Edessa 78 familia publicanorum 212
chartered joint stock companies, Coromandel Coast 145 Egypt 4, 191, 223, 232, 286 famines and trade 77-8
European 42, 149, 207 corpora naviculariorum 247-8 corporate commercial groups in 246 Fanon, Franz 24

352 353

Fayum. the 155
Fei'Euson, Adam 6-t. 301
fides (as symbolic capital) 259-62
see also credit
Finley. M. I. 26, 28, 33, 42, 47, 282
and anthropology 22-4
on food crisis in Antioch 143
on merchant associations
and the politics of otherness 32
.
/i'.S-CK.f 223
Foucault. Michel 24
fragmentation, economic 5, 201, 287
frankincense 134, 202
Frederiksen, Martin 26-7. 29
Gaius (the Romanjurist) 131, 174, 277
Garnsey. Peter 25
Gaul 72-3, 127
Gaza 192
Geertz, Clifforcf 5, 6, 129, 131. 197, 198
see also bazaar
Gellius. Aulus 102
Gibbon. Edward 61
gift-giving 25, [07
Goethe, J. W. von l
Goitein, S. D. 43
Goody, Jack 44
Great Indian Mutiny 80
Gregory of Nazianus 140
growth, economic 48, 62, 122-7
aggregate or per capita
Guadalquivir Valley 77
Gujarat 78, 220
Hadrian 151, 152, 178, 222
Hall, John 66
Hegel, G. W. F. 35
Heliodorus 195
Hercules, cult of 261
Hercler, J. G. t1
Herod the Great 179
hierarchies, urban 35-6, 126-7
high culture, urbanised 6-7
see also agrarianate society
historical sociology 7
hoarding 175-6
Hodgson, Marshall 11, 41
see also agrarianate society
honestiores and humiliores 124
354
INDEX INDEX
Hopkins, Keith 25, -jo. 120 justice, administration of 188-90. 262 imbalances and imperfections 139-40,
and Finley 26. 47 interaction of courts and private 142
and quantification 86 mediation 266-7 modem 4
on the Roman senate 98 mediation 263-5 the Roman state as actor in 176-9
see also taxes and trade rough 203, 240, 264-6, 268 as transfbrmers of agricultural
Horace 303 self-help 262 surplus 119-21, 294
households in the economy 5, 19, 31-2, volatility of prices 4, 136, [42, 166-71
242 78. 241, 268 Kautsky. J. 65 Marx, Karl 35, 45
himdi, the 149 Kipling. Rudyard i measures, variation in standards of 191-2
and cashless transfers of money in Melanesia -i8
Roman society 285-6 La Graut'esenque 134 Melville, Herman 132
laissez-faire 5, 51-4, 140, 239, 291 mercantilism/mercantilist privileges 5,
imperial government, weakness of 68-9, see also market integration; Smith; 51-2. 56, 65, 74, 147, 213
239-40, 262 Ricardo merchant princes 196, 258, 283-4
imperialism, European 40 Lane, F. C. 2^2-1 merchants
Inca Empire 7 see also protection acting as patrons 259
incense 291, 302 Lattimore, Owen 25 bargaining power of 151. 225-9, 248
Indian Mutiny see Great Indian Mutiny Lepcis Magna 50 fiscal monopoly of 152
Indian Ocean 8, 41-2, 52, 53, 76-7, [96, Leptiminus 73 patronage of 228-9
3"4 Levy-Bruhl, Lucien 38 religious association 254-7
trade and the agricultural surplus 76-7 Lex Irnitana 175 smuggling of 225-7
industrialisation 85 Libaniu.s 101, 103, 105 superstitions of 132
institutions 12-13 on the autonomy of the market 175 see also bazaar; markets;
see also Akerlof: bazaar: markets; local knowledge 194, 199 middleman, the
North: transaction costs see also Geertz Mesopotamia, ancient 271
integration, economic 12, 29, ll4-l5> see London 34, 47 Meyer, Eduard 19-21
also empire: tributary, and Lugdunum 253 middleman, the 5, 50-7, 147-8
economic integration: market Luke (the evangelist) 265 Mitteis, Ludwig 184
integration Lyeia 106-9 Moby Dick 132
integration, legal [80, 183-90 modernisation 6
122-3, 125 clash of legal practices 189-90 Maecenas 303 Mommsen, Theodor 17-18
co-existence of Maine, Henry Sumner 18, 25 monopoly, fiscal 246
Egyptian custom 188 Malinowski, Bronislaw 38 Monte Testaccio 35, 273
Muslim and Hindu [aw I 83 Malthusian trap 34 Montesquieu 17, 54-60, 63, 74
Roman and local law I 83 Mann, Michael 122 Mughal Empire 8-10, 79-84
Roman and Jewish law I 85-6, mansabdars 95-6 comparison with Roman Empire 13,
190 Marcus Aurelius 61. 280 57, 292-3
pragmatic application of imperial mare clausum 135 comparison summarised 293-6
law 185, 188-9 market integration 4, 13, 30-2, 68-9, 79, customs regimes 213-25
see also justice, administration of 110-15, 140 elite building 93-110
integration, monetary 180-3 criteria of 154-5, 171-3 imperial styles of
Istanbul 40 limits to in the bazaar 131-201 consumption 296-304
pasSim limits to local
Jewish diaspora 262 and regional variation 157-66 control/penetration 183-90,
Jones, A. H. M. 22, 26, 133 markets 4, 49, 140 206-7, 225-7, 268
Josephus 103-4, 179 and big cities 142-3 market functioning as tribute
Jucundus 121 conditions of trade 111-2 transtormer 119-21
Julian (emperor) 143, 224 equilibrium 140 Mughal bazaars as bench-mark for
Juliopolis 240 and harvest cycle 165-6 Roman markets 149-53
355
 1N D E X INDEX

Mughal Empire (cant.) Paul (the Roman jurist) 278 provisioning policies see consumer Smith, Adam 51, 62, 147, l52, 240
and old world Eurasian trade 304-6 Fax Romana 29, 70 city/interests Smithian dynamics 45, 48, 291
promoting corporate peace dividend 235 publicans <j)iihlicani) 121, 204 see also agrarianate society: economic
associations 246-8 quality of protection offered by 228-9, weak state constraints on 206-12 integration; markets
as tributary systems 110-2 l 238 Puteoli 253, 272. 283-4 societas 275-6
Murecine wax tablets see Sulpicii peculium 276 Tyrians in 254 and commenda 276
archive/bankers and limited liability 277-80 see also partnership
Muziris 281, 281-4 see also commenda: credit Rabirius Postumus, Caius 142 Sombart, Werner 299
pedlar-market 196-7 Red Sea 254, 281 Song dynasty 43
Navigation Act 56 see also bazaar redistribution 67-9, 72-3, 119-21, sources see evidence
Needham, Joseph 43 Peking see Beijing 125-7 Spain 73
Nehalenia 1^1. 261 Pelsaert, Francisco 145, 199 revenues, prebendal 96-7, 125 Spasinou Charax 253
Nero (attempt to curb abusive conduct of see also chartered joint stock Rhodes 231 spices 302
publicans) 207 companies Rhodian Sea Law 191 state

Nerva 98 pepper 146 Rhone-Saone-Rhinearterial route 71, segmentary 111

networks of trade, Pergamon 216 [27 size of Mughal 83-4, 93-4
compartmentalised 287-9 Petronius 132, 264-5, 301 Ricardo, David 72 size of Roman 69-70, 92-3, 115-16
Nile, River 155, 281 Philo 245 see also economic integration state-system, European 55-6
North, Douglass 49, 140 Piraeus 151 risk and uncertainty 4, 13 Stein, Burton 111
see also Akerlof: transaction costs pirates 131, 203, 230-2, 237 in trade 132 stock exchanges 50
Pisa 288 and the bazaar 199, 240-1, 269, Sulpicii archive/bankers 187-8, 267-8,
Obasanjo, Olusegun 38 see also terra sigillata 287 272, 279, 283-4
Obeysekere, Gananath 37 Plato 262 risk, diversification of 195-20 l, 275 Surat 57, 196, 243, 244, 247
Oenoanda 109-10 Plautus 277 Rodinson, Maxime 43 surplus extraction 78-9
olive oil 73-6, 289 Pliny (the Elder) 134, 135, 202, 302 Romaia Sebasta. festival of 216 see also taxes and trade
see also Monte Testaccio Pliny (the Younger) 99, 139, 240 Rome (the imperial capital) 2, 9, 142, Syria 66
Opium Wars 58 Plutarch 103, 105 222, 247, 254, 280
Opramoas of Rhodiapolis 106-9 Pol Pot 24 compared to London 47 Tacitus 98, 99, 207
Orient/Orientalism 1-2, 44 Polanyi, Karl 23, 49, 67, 119 concentration of consumption in 47, TajMahal 80, 296-8
and classics 17 see also anthropology; Finley; Geertz; 66, 127 Tarsus 248
Oriental luxury goods 121, 258, 281, Malinowski; primitivism- the grain supply of i 76-8, 273 Tavemier, J.-B. 96
305-6 modemism; redistribution; Rostovtzeff, M. I. 21, 29-55, 74, 75, 258, tax-t'armers
and customs dues 214-15, 228 Sahlins 291 see publicans
Ostia Pollux 202-3 taxation

Piazzaledelle Corporazioni 250-3 Pompeii 121 Sabora 223 and price records 155-6
"othemess", the politics of 6, 23, 32-3, population size 86, 124 Sahlins, Marshall 37, 290, 302 Roman and Mughal 79
37-40, 43-4, 46 predation 12, 13, 62, 66, 83, 122, 150-1, see also anthropology; "otherness", the see also custom dues; tribute
Oxyrhynchus 166, 248 204-12, 226, 235, 239 politics of taxes and trade 8, 12, 47-50, 62, 83,
primitivism-modemism I, [2, 19-36, Said, Edward 44 113-21, 291, 293-4
Palmyra 151, 196, 221-2, 228, 230-1, 291-2 Sardis 248 Tebtunis 170
283-4, 305 parallel debates 8, 37-45 Sartre, Jean-Paul 24 temple of Bel 254, 257
customs law of 224 private (aristocratic) property 97-110 Scaevola 280 terra sigillata 30, 74, 288-9, 304
merchant community of 253-4, production, decentralised 30, 74-6 Seleucia 254 Tertullian 245
257 productivity gains 122-3, 125 Seneca 101 theory
Pamuk, Orhan 40 protection 13, 200-1 seviri Auguxtales 259 see conceptualisation
Panopolis 68-9, 79 costs of 233-4 Shahjahan 80 third world, the 6
partnership 270, 273-6 locally organised 230-2, 234 Sidonius 143 Timur Lenk 79
see also commenda and predation 203-4 silver, American 306 Titus 103
Parthian Empire 18 see also Fax Romana slavery 274, 276-80 tonnage, ships 136

356 357
 INDEX
II trade van Leur, J. C. 41-2, [29, 28l
European see capitalism Venice 57, 305
irregularities of 5, 12-13, 131-201; see Verres. Gaius 106
also risk and uncertainties Vespasian 103, 223
seasonality of 138 Veyne, Paul i, 25, 107, 214, 301
see also bazaar, markets Vindolanda 151, 152
Trajan 98, 99, 240 violence, endemic 102, 229-30
transaction costs 140, 197, 202-7, Vologesias 253, 254
219-20, 226, 229-38, 286-7, 289 Voltaire 56, 63
see also Akerlof; North
transport 133-6 Weber, Max 3-4, 7, 17, 23, 32'47.
tribute 52, 53, 62 67
and curtailment of fiscal on the ancient economy 20-1
rebates 216-18 on commenda 271
and rent 10. ill on political capitalism 122
and Roman and Mughal elite on the politics of othemess 33
building 93 and tributary empire 63-5
see also empire: tributary, and economic weighingand filling, as negotiable
integration process 192-3
Winckelmann. Johann Joachim 19
Ulpian 203 wine 73-6, 166-71
Ulpius Yarhai, Marcus 196, 258 Wolf, Eric 41
underemployment, rural 124 world-system 5, 76
urbanisation. ancient and early
modern 34-6 Zamindars 94
see also consumer city/interests Zheng He 52-3

358